Exploring the implementation of the school-based obesity prevention Familie⁺ intervention: a pilot mixed-methods study of teachers’ perceptions

Abstract

Teachers play a pivotal role as implementers in school-based obesity prevention interventions. Given the influence of teachers’ perceptions on the implementation process, the aim of this pilot study was to investigate the delivery of, ratings on, and influencing factors for the implementation of the school-based Familie⁺ intervention delivered in low socioeconomic regions in Germany. Building upon the Consolidated Framework for Implementation Research and the Implementation Outcomes by Proctor, we conducted a sequential mixed-methods study design. Delivery of the intervention activities, their ratings, and influencing factors were assessed using questionnaires and interviews at two time points. Deductive qualitative content analysis using Consolidated Framework for Implementation Research was applied to the interview data to support the questionnaire data. Mixed-methods data are jointly presented. Overall, delivery of the intervention varied, with the nutrition activities most often delivered. Teachers valued the intervention activities overall positively. The main factors influencing implementation were related to the intervention materials themselves (e.g. high complexity and limited adaptability) and procedures within the school (e.g. mixed compatibility with workflows). School-based interventions should incorporate principal engagement and strengthen cooperation within the school team. Flexible interventions with multiple options for adaptations should be provided.

Background

Childhood overweight and obesity are major public health concerns [1]. Children with overweight are likely to become overweight as adults too [2], increasing the risk of developing health consequences, such as diabetes or cardiovascular diseases [1]. This burden is especially high in groups with a lower socioeconomic status (SES) (e.g. low educational level or low income), as there is a higher prevalence of obesity and overweight in those groups compared with groups with a higher SES [1, 3].

Schools are an important setting for obesity prevention, reaching all children, regardless of their SES, over a prolonged period of time [2, 4, 5]. There is moderate evidence supporting the beneficial effects of multicomponent interventions addressing obesity prevention in the school setting [6–8]. Furthermore, it is known that implementation processes influence the effectiveness of interventions [9–11]; for example, a higher dosage implemented is associated with better health outcomes [9, 10].

Within the school setting, teachers often play a pivotal role in the implementation process of health-promoting interventions [12]. Teachers’ implementation efforts are influenced by their characteristics [13, 14] and beliefs and attitudes towards the intervention [13, 15]. Furthermore, to effectively take up their role as health promoters, teachers need principal and collegial support [16, 17] and knowledge and skills [18]. Previous studies conducted mixed-methods analysis of teachers’ perceptions [19, 20], some specifically with a focus on nutrition and/or physical activity (PA) interventions [15, 21, 22]. Applied methods included for example convergent mixed-methods design guided by implementation science frameworks, merging data at the results level [19], Bergling et al. analysed qualitative and quantitative data separately before identifying meta-inferences across congruent and incongruent findings [15], and examining determinants of sustainability by grouping participants based on their perceived likelihood of program continuation [15, 21]. These studies found organizational factors such as school priorities, culture, and leadership to be more influential than personal beliefs of teachers [15, 16]. For example, enough time for intervention delivery was found to be an important factor for implementation [15, 23]. More teaching experience led to more intervention components delivered [22], as well as an adaptable intervention [19], and if the interventions seemed to meet students’ needs [15, 19].

Gaining a deeper understanding of teachers’ perceptions and experiences as implementers is essential [11, 15] in order to incorporate those perceptions into implementation processes of interventions [24, 25] and contribute to their intentions for sustainment [26] and institutionalization [27] of interventions. The sustainability of school-based interventions, for example, is influenced by staff engagement and access to resources and equipment [16].

An example of a comprehensive obesity prevention intervention targeting predominantly inhabitants with low SES, consisting of three sub-interventions (i.e. primary school, family, and community intervention), is the Familie⁺ project, which builds on collaborative relationships with participating schools. We conducted a pilot study in three regions in Germany to evaluate the school-based sub-intervention of Familie⁺ and its implementation processes. The aim of this pilot study was to evaluate teachers’ perceptions of the implementation process and intentions to sustain the school-based Familie⁺ intervention using a mixed-methods approach.

Therefore, the following research questions (RQ) were addressed:

1. How was the delivery of the school-based Familie⁺ intervention?

2. What were teachers’ perceptions on the acceptability, feasibility, and appropriateness of school-based Familie⁺ intervention?

3. What factors influenced the implementation of the school-based Familie⁺ intervention?

4. How do teachers' perceptions differ between those with and without the intention to sustain the intervention?

Methods

Intervention description

The overall Familie⁺ project aimed to promote health behaviours that reduce the risk of obesity. Its goals included increasing PA, improving healthy nutrition intake, reducing sedentary behaviour, and enhancing both the duration and quality of sleep [27]. The Familie⁺ project consisted of three sub-interventions: a (1) school- and (2) family-based intervention designed to address the same aims across both settings, and (3) a community intervention with the aim to build/strengthen the network on child health promotion in the respective community. This study focuses only on the implementation evaluation of the school-based intervention.

To develop the school-based intervention, we conducted a comprehensive review of existing school-based interventions and the available evidence base in German-speaking countries. The best available intervention materials addressed children between 8 and 12 years and were used to inform 45 intervention activities. During sessions with the project team consisting of researchers, teachers, health promotion practitioners, and collaborating schools, the intervention activities were developed. In the intervention, teachers were intervention deliverers and facilitators who distributed intervention activities to families. This involved sending home letters with students, ensuring that families receive the corresponding materials for each intervention activity.

An in-person voluntary workshop by a health scientist and a teacher from the project team to prepare the teachers for the implementation of the intervention was offered to all schools. Four out of nine schools participated in this workshop, lasting 1.5 to 3 h. Intervention activities were provided during the workshops or sent via post; a digital version was also provided. The four health behaviours, PA, nutrition, sedentary behaviour, and sleep, were broken down into eleven smaller subtopics. For example, for improving healthy nutrition intake, three subtopics with specific aims were addressed: Promoting water instead of sugar-sweetened beverages, promoting the consumption of fruits and vegetables, and promoting healthy snacks instead of unhealthy snacks. Multiple intervention activities were provided for each subtopic. Teachers were asked to address all eleven subtopics by delivering at least 11 intervention activities. They were free to choose the intervention activity within the subtopics. As teachers are autonomous in managing their classes and lessons, there were no specific requirements regarding when to implement specific intervention activities. Intervention activities included pre-prepared lessons (including working sheets for students and background information on the lesson topic for teachers), class activities (e.g. vegetable card game, PA protocol), and school activities (e.g. guide on how to increase the movement friendliness of the schoolyard, installation of water dispensers). Although the intervention was initially planned to last 11 weeks, with one topic each week, starting in September 2021, the COVID-19 pandemic necessitated a flexible implementation approach, allowing schools to determine their own start dates. Throughout the implementation of the intervention, teachers were offered ongoing support by the research team through regular telephone calls and emails (for further intervention description following the TiDIER Checklist [28], see Supplementary File 1).

Study design for this investigation

This pilot study has a sequential mixed-methods study design [29], allowing for data assessment with breadth and depth. Data collection started with quantitative data, gathered sequentially with teacher questionnaires at two time points (08/2021 with the start of the intervention and 03/2022, 3–10 weeks after the intervention ended), after teachers gave informed consent. Subsequently, qualitative semi-structured interviews were conducted to provide further in-depth understanding. The design leans towards being partially mixed since two of the four research questions are answered solely through quantitative data, indicating that quantitative data are dominant.

Conceptual framework for this investigation

This pilot study was guided by the Consolidated Framework for Implementation Research (CFIR) [30, 31] and the Implementation Outcomes by Proctor [32, 33] with corresponding German translations [34, 35] for the development of measurement instruments and analysis procedures.

The CFIR is a comprehensive and commonly used framework on influencing factors for implementation, applicable across divers settings and comprising five domains: (1) intervention characteristics include all aspects related to the implemented ‘thing’ (e.g. intervention), (2) inner setting pertains to features of the setting where the intervention is implemented (e.g. schools), (3) outer setting refers to the broader context in terms of structural, political, and cultural aspects, (4) characteristics of individuals include traits of individuals involved in the intervention (e.g. teachers), and (5) process covers all strategies and processes involved in implementing the intervention [30, 31]. We combined this with implementation outcomes defined by Proctor [32]: acceptability, adoption, appropriateness, costs, feasibility, fidelity, penetration, and sustainability. Sustainment is also called maintenance and can be defined as the continued implementation of the intervention after external support ends [36].

The quantitative questionnaires were designed based on the first version of CFIR [31]. At the same time, the coding of the qualitative data was aligned with the updated version [30], which was not available at the time of the start of the study.

Participants

During the school year 2021–2022, the Familie⁺ intervention was implemented in nine primary schools located in three predominantly low SES regions in Germany. As the intervention targeted children between 8 and 12 years, it was decided to address third and fourth (students aged 8–11). All 47 teachers from the third and fourth grades were invited to participate in the evaluation.

There are no standardized and freely available data on SES at the school or even municipal level in Germany. The selection of the recruiting areas within the municipalities was based on various administrative data, serving as a proxy for SES [37]. Those included information on income, the number of children under 14 in families receiving social security benefits, the percentage of individuals with a migration background, and the prevalence of obesity. In cooperation with the local public health experts, schools in the recruiting areas were contacted, and nine primary schools could be recruited for the Familie⁺ project.

Instrumentation and data collection procedures

Data were collected at two different time points, during (inter) and after (post) the intervention implementation phase. During the intervention (inter), teachers were asked to document each intervention activity they delivered using a concise one-page paper questionnaire (Supplementary File 2). We assessed delivery of the intervention (RQ1): the extent to which teachers delivered the intervention activities, and if it was delivered as intended (per protocol, adapted, or not finished). Ratings on teachers’ acceptability, perceptions on student acceptability, feasibility, and appropriateness were assessed (RQ2) with one self-developed item each, on a 5-point Likert scale (‘strongly disagree’ to ‘strongly agree’; ‘very good’ to ‘very bad’). The items were developed to fit the specific context based on Gutt et al. [35] and were reviewed in consultation with a psychologist. After the intervention implementation (post), teachers received a subsequent online questionnaire (Supplementary File 4) addressing teachers’ perceptions again, this time on the whole school-based intervention. Teachers’ ratings on acceptability, perceptions on student acceptability, feasibility, and appropriateness (RQ2), as well as intention to sustain using the intervention (RQ4) (‘Are you planning to continue using the Familie⁺ materials?’) were assessed. In addition, CFIR influencing factors for implementation (RQ3) were assessed using self-developed items based on Regauer et al. [34]. To keep the questionnaire to a reasonable length, influencing factors for implementation were selected beforehand in discussions within the project team to be seen as most relevant to schools and teachers about the intervention (see Supplementary File 5). Three questions were added, assessing the impact of the COVID-19 pandemic on the implementation, students’ health, and communication with parents. All assessed aspects were presented on a 5-point Likert scale (‘strongly disagree’ to ‘strongly agree’).

In addition to the questionnaires, all teachers were invited to participate in semi-structured interviews (Supplementary File 3) to assess the implementation (RQ3) and influencing factors for implementation (RQ4) of the intervention via video calls (post) with a duration of 45 min on average. Table A in Supplementary File 5 provides an overview of all assessed influencing factors in the (post) questionnaire and the interviews.

Data analysis

Questionnaires

Descriptive analysis (RQ1), including means and standard deviations, was calculated for each intervention activity and across all activities (inter) and for the whole intervention (post) for each rating on teachers’ and students’ acceptability, feasibility, and appropriateness. For calculating the ratings’ values, 1 to 5 were used (‘strongly disagree’ to ‘strongly agree’; ‘very good’ to ‘very bad’). In addition to the means, ‘agree’ and ‘strongly agree’ (‘4’ and ‘5’) were summed up to evaluate the percentage of teachers with positive ratings on the intervention (RQ2). For evaluating the influencing factors for implementation, 1 to 5 were used (‘strongly disagree’ to ‘strongly agree’) to calculate frequencies of the disagreement (‘1’ and ‘2’), neutrality (‘3’), and agreement (‘4’ and ‘5’) with influencing factors (RQ3). Quantitative data from questionnaires were analysed using Microsoft Excel 2016 [38].

For an explorative group comparison [33], teachers were dichotomized into a ‘yes’ and ‘no’ group. In a second step, we compared the ratings of the no intention to sustain to the intention to sustain group for the overall school-based intervention and the influencing factors (RQ4).

Interviews

Interviews were transcribed verbatim and analysed in MAXQDA 2020 [39]. All data were coded by F.B. and M.S. separately by applying qualitative content analysis. The procedure of content-related structuring was applied to the data following the CFIR constructs and outcome addendum [33] as deductive category system. In regular meetings with co-authors (F.B., M.S., D.G.), the procedures of assigning codes to segments was discussed for consensus [40].

Mixed-methods analysis

Data merging was applied after the questionnaires and interviews were analysed separately [40]. For the overall rating of the intervention (RQ2) and for influencing factors for implementation (RQ3), descriptive measures from the quantitative data assessment are jointly presented with results from the qualitative data analysis. Summaries of the qualitative data and a quote from the material are presented together with the quantitative results from the corresponding influencing factor.

Results

From the 47 teachers invited to participate in all the data assessments, we received 22 inter-questionnaires (47%) and 18 (38%) post-questionnaires. Seven teachers (15%) agreed to participate in an interview. Six teachers (13%) participated in all three data assessments from our sample. The sample as participant flow diagram and as number of participants for each data assessment is presented in Fig. 1. Known reasons for drop-outs were discontent with intervention activities, other reasons for drop-outs and drop-in were changes in staffing.

Figure 1.

Illustration of the sample by (a) participant flow diagram and as (b) number of participants for each data assessment.

Delivery of intervention activities

The teachers (n = 22) delivered 162 intervention activities in total, at all schools during the whole intervention phase, with an average of 7.4 [1, 19] activities per teacher. Of the 162 activities, information on delivery was available for 157. Among these, 98 activities (62%) were delivered as originally described, 57 (36%) were adapted in some way, and 2 (1%) were not delivered. (Supplementary File 6 presents all delivered activities). Of the 45 intervention activities available, 38 (84%) were delivered at least once in any of the schools. As shown in Fig. 2, nutrition was the most frequent delivered topic with 89 activities (55%), followed by sleep with 32 activities (20%). Of the delivered intervention activities, 117 (72%) were lessons, 43 (27%) class activities, and 2 (1%) were school activities. These results show variability in delivery between teachers and therefore in class delivery.

Figure 2.

Delivery of intervention including distribution of (a) topics and (b) formats among delivered activities and distribution of (c) delivered and not delivered activities among available activities.

Teachers’ perceptions

Table 1 presents mean ratings on the intervention activities (inter) and on the whole intervention (post) along with quotes from interviews on teachers’ acceptability, perceptions on students’ acceptability, feasibility, and appropriateness (RQ2). Supplementary File 6 presents all delivered activities (inter) along with their respective ratings.

Table 1.

Joint display of teachers’ quantitative and qualitative ratings of intervention activities and the whole intervention on acceptability, feasibility, student acceptability, and appropriateness.

	Intervention activities average ratings		Whole intervention ratings
	Mean [SD] a	% of teachers agreeing b	Mean [SD] a	% of teachers agreeing b	Quotes from interviews
Acceptability	3.86 [0.94]	73%	3.94 [0.94]	67%	‘I think there were really too many ideas […], but there were many practical things included, which was really great.’ (Interview 03, Line. 111) ‘Therefore, I was actually the only one who did it [the intervention], in the end.’ (Interview 01, Line 138)
Student acceptability	4.04 [0.86]	80%	3.72 [0.96]	61%	‘. . . and the kids said, Oh, this again [nutrition]?’ (Interview 01, line 62) ‘Only the aspect of sleep, that was new for the kids.’ (Interview 06, Line 16) ‘I believe what the kids really liked was the practical connection, you know, to their everyday life with nutrition and PA.’ (Interview 02, Line 153)
Feasibility	3.95 [0.89]	74%	3.33 [0.90]	56%	‘I always managed only about half of what was planned for an [school] hour.’ (Interview 05, Line 91)
Appropriateness	4.11 [0.90]	84%	3.83 [0.62]	72%	‘It was very detailed and comprehensive, so I think even my very committed parents eventually stopped reading it.’ (Interview 05, Line 56)

^aOn a 5-point score, where 5 is the fully agreement with the intervention being acceptable, feasible, appropriate.

^b% of teachers agree/fully agree [4, 5] with the intervention being acceptable, feasible, appropriate; SD = standard deviation.

With the exception of the mean ratings on teacher acceptability, all ratings and percentage of teachers agree/fully agree slightly decreased comparing the ratings of the intervention activities (inter) and of the whole intervention (post).

During the interviews, teachers (n = 7) expressed that the intervention was overall too extensive, however, it consisted of valuable and new ideas, even for an experienced science class teacher. Teachers expressed their view that the activities designed for parents were also overly extensive and presumed that even motivated parents did not read them.

Opinions among teachers differed regarding the topics addressed. On the one hand, some teachers mentioned that their students were already familiar with all aspects of PA and nutrition, resulting in a lack of interest. On the contrary, teachers stated that students could relate to those topics, making them more relevant. For sedentary behaviour and sleep, some teachers considering these topics as novel and interesting, while others believed that students were not engaged in those behaviours, and therefore, are not interested.

Delivering entire activities posed challenges, because the planned time and scope were too optimistic and too extensive. One teacher reported being the sole person involved in the intervention implementation process at school, highlighting a lack of participation from others and even destructive behaviour towards both the research team and the school team in relation to the intervention.

Factors influencing implementation

All influencing implementation factors assessed in the questionnaire (post) and the interviews are presented in Supplementary File 5. Results from the questionnaire (post) is presented in Fig. 3. In the following results from the questionnaire, the interviews are jointly presented.

Figure 3.

Teachers’ perceptions of factors influencing implementation.

Intervention activities

Some intervention activities were perceived as too extensive and therefore too complex for implementation, though others were perceived as clear and straight forward. ‘I just don't have the time to go through such a huge thing entirely. […] but there are some great ideas in there.’ (Interview 05, Line 81). Half of the teachers (50%) perceived the activities as too complex.

The adaptability of the activities was limited according to teachers. ‘I don’t know if I would agree, that the Familie⁺ activities were especially well designed for a differentiation between different proficiency level of the students’ (Interview 01, Line 50). However, during the post-questionnaire, 67% of teachers agreed that the activities were adaptable to the needs of their students, whereas 28% agreed that activities were adaptable to context, and 39% were neutral regarding the adaptability to context.

One positive aspect about the intervention was its layout. ‘I found it appealing.’ (Interview 06, Line 28). This was supported by the questionnaire data showing that 78% agreed that the design was good. Overall, 33% of teachers perceived no advantage of Familie⁺ compared with other similar health projects, while 50% were neutral regarding an advantage.

Procedures within school

In total, 78% of the teachers completing the questionnaire agreed that their head teacher supported implementation ‘We need to stick to their [head teacher] vision, but then the head teacher really supports us’ (Interview 02, Line 26).

Half of the teachers (56%) agreed that they communicated about the Familie⁺ intervention within the school team. Teachers valued additional implementation efforts of other school staff, like the afternoon care personnel offering fruits and vegetables or a social worker at school, though the support was regarded as too little. ‘We have a social worker once a week for 3 hours, so she can’t do a lot’ (Interview 07, Line 107–112).

Procurement of materials in school was perceived as complicated due to the bureaucratic procedures and the limited copy quota. ‘There are always discussions on how much we are allowed to copy.’ (Interview 02, Line 127). There was some budget provided through the research project, which was mentioned as helpful. Based on the questionnaire data, 39% of teachers agreed that the financial resources were sufficient for conducting the intervention, 39% were neutral whether financial resources were sufficient. In total, 78% disagreed with the statement of having sufficient personnel. Covering lessons of sick colleagues increased teachers’ workload, but was also perceived as common practice of being a teacher. ‘Substitute lessons will be daily business on the next 10 years’ (Interview 01, Line 98).

Regarding the delivery of intervention activities, the compatibility was perceived differently. ‘Many lessons were prepared in a way that they could be delivered in one school lesson’ (Interview 01, Line 42), 56% of teachers agreed that Familie⁺ was compatible with their workflow and 39% agreed that Familie⁺ was compatible with the curriculum. ‘It just doesn’t fit in with the other topics in the subject’ (Interview 05, Line 50–52).

It is important to have someone, who is passionate about it and leads the process according the teachers: ‘he or she must then involve others, then a system can begin to change’ (Interview 04, Line 70). Most teachers (88%) stated that there was a person in their school, who was passionate about the implementation process.

The COVID-19 pandemic led to a shift in priorities, as most teachers emphasized the need to prioritize major subjects, like reading, writing, and mathematics. ‘In times of Corona, learning deficits must be prioritized over health education’ (Interview 04, Line 32). On the other hand, teachers mentioned that, especially during the pandemic, health and health promotion, and social competences were regarded as important. All teachers (100%) stated that the COVID-19 pandemic had a negative influence on their day-to-day work, on the implementation of Familie⁺, and on the health of students.

Group comparison intention to sustain using intervention

Figure 4 presents the group comparison of teachers with and without the intention to sustain the intervention, i.e. intention to sustain (n = 13) versus no intention to sustain (n = 7), for positive perceptions (i.e. agreeing/strongly agreeing) on appropriateness, students’ acceptability, teacher acceptability, and feasibility and with influencing factors for implementation.

Figure 4.

Group comparison intention to sustain using the intervention.

Overall, potential non-sustainers scored lower than potential sustainers. The smallest difference between the groups was found in the rating of appropriateness (60% vs. 80%).

Looking at factors influencing implementation, both groups perceived the relative advantage (both 17%) and adaptability of the intervention to their context relatively low (20% and 40%, respectively). Regarding students’ needs sustainers perceived this higher (25% vs. 95%, respectively). The sustainers group rather perceived the intervention activities as not too complex (71% vs. 36%). In total, 43% of the non-sustainers, and all teachers (100%) of the sustainers agreed, that their principal supported implementation. All (100%) in the sustainers group agreed that there was a person who led the implementation process, versus 67% in the non-sustainers group, but available resources was regarded as low in both groups (0% and 27%, respectively).

Discussion

The aim of this pilot study was to evaluate implementation processes and investigate teachers’ perceptions towards the implementation and sustainment of the school-based Familie⁺ obesity prevention intervention. A majority (84%) of the intervention activities were delivered at least once (RQ1), and the ratings were good (RQ2). Influencing factors for implementation (RQ3) were predominantly identified regarding the intervention activities themselves (e.g. complexity and adaptability) and procedures within the school (e.g. procurement of materials, compatibility with workflow). Teachers, who indicated an intention to sustain the intervention, showed more positive perceptions on the interventions’ acceptability and adaptability (RQ4).

Our response rate was lower than expected. Response rates among teachers in evaluation studies in Germany vary considerably, ranging from 33% [41] to 89% [42]. Such variation can be the result of how closely teachers are involved in and perceive the relevance of the evaluated program. In this study, principals decided whether their schools would participate in Familie⁺. Teachers chose whether they wanted to implement the intervention in their classes and, importantly, whether to take part in the evaluation. Other studies have also shown a lower response rate in teachers [43], indicating the evaluation among teachers is challenging. Yet, we also collected interview data, therefore we not only rely on questionnaire data.

Teachers’ ratings of the intervention activities and the whole intervention were quite high. However, interview data revealed more nuanced and sometimes mixed perceptions. Eisman et al. [19] also found good ratings for the intervention and lower fidelity to the intervention than targeted. The wide range of teachers’ ratings indicates individual implementation preferences, which was also stressed during the interviews. Differences in teachers’ perceptions of individual components, combined with additional contextual influencing factors, likely contributed to variability in the delivery of intervention activities.

Among the delivered activities, nutrition was the most frequent delivered topic (55%), followed by sleep (20%). This might be explained by the fact that nutrition was chronologically the first topic within the intervention. An alternative explanation could be that, out of the four topics, nutrition and PA were already integrated into the school curriculum, as stated by teachers. Teachers might have replaced or enriched their usual resources with the Familie⁺ activities. Sleep was seen as a new topic to be addressed in the school setting, explaining the rather high interest in this topic. Reviews on school-based health promotion also show the relatively innovative topic of sleep, with only one intervention in the included articles addressing sleep [16, 44].

Among the delivered activities, 36% were adapted for implementation. This number seems to be low, considering teachers have individual teaching approaches and conditions within their classes [26]. The amount of adapted activities could be low, as activities did fit with existing school workflows, systems, and processes [30], and therefore, the need for adaptations was lower. However, scores for compatibility with the curriculum and workflow varied in our study. Compatibility is connected with the interventions' adaptability, if it can be modified, tailored, or refined to fit local context or needs [30]. Therefore, it is also possible that teachers decided not to deliver an intervention activity, rather than adapting it.

The SES of students could have also played a role in the process of adaptations or non-delivery of intervention activities. Schools in the Familie⁺ project were situated in regions with predominantly inhabitants with low SES, these schools face additional challenges compared with high SES regions [45]. This may be relevant to the schools in this study, as various challenges were mentioned in the interviews to be influencing implementation (e.g. students stealing food from others due to hunger, language barriers with parents). These findings are consistent with a systematic review that points to a wide range of influencing factors for the implementation of interventions addressing students with low SES [44].

As part of the intervention in this study, schools were offered several environment-focused activities. This included, for example, a financial support program (€3 500 per school) for procuring and installing a water dispenser, and guidance on redesigning schoolyards to be more PA friendly. Financial and policy support aligned with intervention activities can facilitate implementation of intervention for children with low SES in school settings [27, 46]. One school initiated such environmental changes; however, the project timeline was too short for full implementation. Therefore, not only behavioural but also structural and environmental changes, together with strong community and policy support are important to ensure long-term effects and address health disparities, but these changes take more time, often exceeding the evaluation period [47].

The sustainment of health promotion interventions is crucial for their long-term effect on health [48]. In this pilot study, therefore, the potential to sustain the intervention was assessed. While the small sample size should be considered, the results from this study are in line with others analysing the sustainment of school-based health promotion intervention and identifying leadership and teamwork as facilitators [16, 26]. A stepwise approach over several years is recommended for better implementation and sustainment [27]. School-based interventions should incorporate engagement of the principal and strengthen cooperation within the school team. Co-production [49] with teachers during the development of interventions can create team spirit, ownership, and make sure interventions fit their context, and strengthening the intention and therefore the potential to sustain the intervention. Principals should be supportive of the time investment for teachers’ training for improving equity-sensitive [50] implementation and sustainment of health-promoting activities [5].

Schools should be understood as complex social systems, and intervention implementation must account for this complexity [51]. Mixed-methods approaches can improve the holistic understanding of context and implementation processes [40]. Implementation should focus on the intervention's core functions while still allowing flexibility regarding the best fitting form. This flexibility enables schools to adapt the activities to their unique local context, resources, and priorities [52]. For example, major differences in proficiency level among students in one class mentioned in the interviews, requires flexible, adaptable intervention activities [53]. Ultimately, interventions can only have be effective, when integrated into existing systems, analysing the effect of single components would therefore be too short-sighted [54]. Balancing this fidelity and adaptation process in public health initiatives remains a challenge, for future research and practice.

More insights and experiences in strategies to support the sustainment of school-based programs is needed [55]. Integrating focus on sustainability and equity within a complex systems framework is suggested but requires further empirical studies [56]. Such an approach is essential not only for sustainability but also for scaling-up and informing policy to improve public health and health equity.

Limitations

This pilot study was conducted in real world conditions with all its challenges, though the following limitations should be considered: All results should be interpreted with regard to the COVID-19 pandemic context, as every school and teacher faced different challenges to different extents at various time points during the implementation period, therefore the intervention period was extended. Although this was a pilot study, the small sample size and the variation in participants across time points represent a limitation. However, the mixed-methods approach helps to partially offset this limitation by providing a more comprehensive understanding of the implementation through different data sources [23]. Assumingly, there is a selection bias of teachers answering questionnaires and willing to participate in an interview. Those who were more positive about the intervention and the Familie⁺ project might have participated more often. This is probably also the reason for only having one interviewee from the group of teachers with no intention to sustain using the intervention activities. Therefore, the group comparison was conducted with quantitative data only. A selection bias can also be assumed regarding the intervention activities, as only activities are rated, that were (partly) conducted.

The analysis for this article was conducted with the school-based data only; a joint analysis with data from the family and community intervention of Familie⁺ could offer a more holistic perspective.

Conclusion

This pilot study contributes to understanding teachers’ perceptions of the implementation of a school-based intervention and shows that teachers should be incorporated when planning an intervention and its implementation. Influenced by school procedures and the intervention itself, teachers’ perceptions and, consequently, their implementation efforts greatly differ. To strengthen sustainability, teachers’ intention to continue the intervention should be addressed from the start. Providing flexible interventions that allow contextual adaptations and diverse delivery formats should be provided and evaluated in order to improve implementation and lastly the health of children. Future research should further explore implementation of such interventions with equal attention to sustainability and equity within complex school systems.

Author contributions

Friederike Butscher (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Visualization, Writing—original draft, Writing—review & editing), Doris Gebhard (Formal analysis, Methodology, Validation, Writing—review & editing), Monika Singer (Data curation, Formal analysis, Validation, Writing—review & editing), Femke Van Nassau (Methodology, Supervision, Validation, Writing—review & editing).

Conflict of interest: None declared.

Funding

This work was supported by the German Federal Ministry for Health: Familie⁺ Grand ZMVI1-2519KIG006.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author upon reasonable request.

Ethics approval and consent to participate

This study was approved by the Ethics committee of the University of Konstanz (41/2021, 08.11.2021) and individuals provided consent prior to participation.