To identify the perceived key elements of the successful Zwolle Healthy City approach, a qualitative study was performed. The study protocol was approved by the Medical Ethics Review Committee (METc) of VU University Medical Center, registered with the US Office for Human Research Protections (OHRP) as IRB00002991. The METc, chaired by prof. dr. C. Boer, confirmed the study did not require medical ethical approval under Dutch legislation on medical trials (2018.601). All procedures performed were in accordance with the WMA declaration of Helsinki.

Semi-structured interviews were carried out with professionals from different organizations and backgrounds into their experiences with the implementation and execution of the approach during the period they were involved (during the entire period of 2010–2018, or part thereof).

The organizations involved are the local government, the regional public health service, a university of applied sciences, two welfare organizations, a home care organization and a municipal sports service. After analyzing the data of the semi-structured interviews, a focus group discussion with a selection of the respondents was held to verify the conclusions. See S1 File for more information on how COREQ requirements were met.

The central approach in this study is ‘Zwolle Healthy City’, an integrated community-based approach aimed at reducing socioeconomic health inequalities, which has been implemented since 2010 in the medium-sized municipality of Zwolle (around 125,000 citizens) in a rural area in the northeast of the Netherlands. In Zwolle, 41 percent of the households form the lowest income group, which is comparable to the country’s average (40 percent) [25]. 80 percent of the adults inhabitants in Zwolle experience their health as ‘(very) good’, 18 percent experience it as ‘okay’ and 2 percent as ‘(very) bad’, which is also comparable to the national averages. 46 percent of the inhabitants that experience their health as ‘(very) bad’ have trouble making the financial ends meet.

In 2005, the Dutch government stimulated that cities would focus on preventing and counteracting overweight in children aged 0–19 years (in the context of the ‘Large cities Policy’). In the municipality of Zwolle, this led to a community-based approach in two disadvantaged neighborhoods. The percentage of households that form the lowest income group are in those neighborhoods 60 and 65 percent. From 2006 through 2009, many health-promoting activities were implemented in both neighborhoods, without structural local coordination.

In 2010, the municipality of Zwolle decided to tackle the prevention of overweight in children in a more integrated and systematic way, inspired by the EPODE approach developed in France [21]. In the Netherlands, this approach has been called the JOGG approach since then. The approach, focused on improving the lifestyles ‘healthy nutrition’ and ‘healthy PA’, was implemented in the municipality of Zwolle, building on the existing community-based approach in the disadvantaged neighborhoods, with some minor adjustments appropriate to the local context. This Zwolle Healthy City approach focused on preventing overweight in children aged 0–19 years living in two disadvantaged neighborhoods.

Since 2014, the Zwolle Healthy City approach has extended its focus, content-wise as well as with respect to the target population, by also including senior citizens [26]. The two disadvantaged neighborhoods were extended with three additional areas. Today, the Zwolle Healthy City approach stimulates healthy behavior by directing towards a healthy lifestyle, healthy social & physical environment and healthy care in disadvantaged neighborhoods, thereby aiming to reduce socioeconomic health inequalities (see Fig 1 for the logic model). Examples of what the approach constitutes content wise can be found in S3 File.

Respondents were purposively sampled to create a heterogeneous interview sample (n = 29, 12 male, 17 female). Purposive sampling is a goal-oriented sampling method in which the respondents are chosen based on the judgment of the researchers (LW, IB) [27]. Respondents were selected based on the organization they work for, their occupation level and their contribution to the implementation of the approach in the years 2010–2018.

To recruit respondents at the tactical level of involved organizations, all members of the program group were invited to participate. The program group consists of professionals working for the organizations involved as manager, policy adviser or (associate) professor. They meet monthly to share and discuss the implementation of the approach. All twenty (previous) members of the program group since 2010 were invited to participate in the study; seventeen of whom agreed to participate in an interview. Reasons for non-participation were (a) the feeling of not being able to contribute, (b) not feeling able to remember much and (c) personal circumstances. To collect a wider perspective on the implementation, the program group members were asked which professionals were involved in the approach at the operational level as sports club advisor, community sports worker, child worker, youth worker or community worker. Based on this, eleven community workers were asked to participate in an interview; eight of whom agreed to participate. Reasons for non-participation were (a) emigration and (b) the feeling of not being able to contribute. Additionally, six municipal executive councilors who had been involved in the approach since the start in 2010 and work at the strategic level, were asked to participate in an interview; five of whom agreed to participate. The reason for non-participation is unknown. All respondents were invited via email (by LW) and, in case of no response, contacted by telephone.

Table 1 lists the organizations the respondents work for and how many of them have occupations at the strategic (n = 4), tactical (n = 17) and operational level (n = 8). The average number of years that respondents were involved in the approach is 5.6 years in the period 2010–2018.

Respondents were interviewed (by LW) in the period from January to March 2019. In a one-to-one interview, they were asked about their experiences with the Zwolle Healthy City approach. The duration of the interviews ranged from 49 to 92 minutes. The last four interviews did not reveal new information, which meant data-saturation was reached. Written informed consent was obtained from the respondents beforehand.

The interview guide was drawn up on the basis of the logic model of the Zwolle Healthy City approach (Fig 1), literature on community-based approaches and documented conversations with involved professionals in 2013 (partly the same professionals). The interview guide can be found in S2 File. The guide consisted of open-ended and follow-up questions structured into the following topics: local structure and coordination, support, interventions and activities, aligning prevention and healthcare, monitoring and evaluation, collaboration, and citizen participation. The interview guide was discussed within the project research team (LW, IB, AJS, JCS, CR).

After analyzing the data of the semi-structured interviews, 7 of the 29 interviewed professionals participated in a verifying focus group discussion. Again, these respondents were purposively sampled to select a representative group for the complete sample in terms of organizations and occupation levels. Four of them work at the tactical level of involved organizations, three at the operational. Respondents were asked to reflect on the report of the key elements that were extracted from analysis of the semi-structured interviews. The respondents agreed with the preliminary conclusions and only had a few suggestions for minor adjustments on the text were made afterwards. These adjustments were made to align the conclusions with the experiences of the respondents. The adjustments did not change the content of the conclusions, but only structure and language.

The interviews were audio recorded and transcribed verbatim anonymously. MAXQDA 2018 was used to analyze the data. Data was analyzed in a systematic way, using the thematic analysis approach with the aim to identify returning themes [28]. To ensure the reliability of data interpretations, interviews were open coded separately by two researchers (LW, DW), by labelling fragments with descriptive codes. Differences were discussed and agreed upon. After open coding five interviews, axial coding took place by creating codes that reflect multiple fragments and codes were grouped together in themes into a coding scheme. The project research team (LW, IB, AJS, JCS, CR) discussed the coding scheme. The coding scheme functioned as a base for the analysis of the following interviews. The coding scheme and themes were regularly discussed by two researchers (LW, DW). Selective coding took place by constantly comparing codes and finding correlations between codes. The derived themes formed the base of the key elements.

According to the respondents, it is important to collaborate with a varied, cross-domain group of organizations that, despite their differences in e.g. mission, tasks and methods, have a common goal. This variety in organizations is typical for an integrated community-based approach that focuses on creating a supportive social and physical environment that enhances healthy behavior from multiple levels and settings. Organizations that can and want to do something for the health and well-being of people in the neighborhood should be involved, e.g. care and welfare organizations, municipal health and sports services and the local government.

According to the respondents, it is helpful that these organizations operate as an independent network that decides its own agenda, in addition to executing ‘basic tasks’ from the local government that most of the involved organizations have. This gives the program group a sense of autonomy. At the same time, two remarks can be made here. Firstly, some respondents indicate that there is a risk of false equivalence in the network because of the dual role that the local government has: as an equal network partner and as the commissioning party. This can lead to confusion about roles and independence, which does not benefit collaboration. Secondly, one respondent indicates that due to the non-hierarchical structure, a sense of a clear task is sometimes lacking for the members of the network, which can be paralyzing. This means that some guidance is desirable.

Furthermore, in addition to collaboration at the tactical level in a network to align practices (the program group), it is important that multiple organizations also collaborate at the operational level. According to the respondents, professionals from different organizations should run projects together and use each other’s expertise.

Respondents describe three levels of all involved organizations that need support for the integrated community-based approach. Firstly, the Zwolle Healthy City approach should be widely supported at the strategic level within the local governmental organization (broad political support) and at the strategic level of other involved organizations. In the past eight years of implementation, municipal executive councilors from the public health domain and other domains (e.g. sports, urban planning and the social domain) were involved in the approach and thereby work towards a healthy city.

Secondly, to gain support at the tactical level from the organizations involved, it is important to act according to the mutual gains approach: not only should the organization’s involvement be meaningful for the approach, but there should also be a ‘win’ for the organization itself.

Thirdly, support for the Zwolle Healthy City approach at the operational level is important among the professionals who run the projects and are in contact with the citizens on a daily basis. One respondent said it is important to have good communication between the professionals working at the tactical and operational level about the goals of the approach and what those goals mean to the professionals in their daily work. Knowing why tasks must be performed and what the relevance is for their work, contributes positively to the support and motivation among professionals to achieve goals that are set.

In addition to these professionals’ need to be involved in decisions made at a higher level, it is important that they experience support to encourage healthy behavior. This support must be more than just the commitment in available working hours. It also involves being given the opportunity to develop necessary skills and to experience moral support from supervisors and other colleagues. Encouraging healthy behavior among citizens must be considered important and must be supported and facilitated by the organization, according to the respondents.

Respondents indicate someone should be appointed to coordinate the network. In the Zwolle Healthy City approach, this is the task of the program manager who, inter alia, organizes meetings for the program group and reminds the organizations involved of their responsibilities with regard to the goals of the approach. According to the respondents, it is, therefore, important that the role of the program manager is independent of the organization he or she works for. The respondents see the fact that the program manager of the Zwolle Healthy City approach is not employed by the local government as a success factor of the approach.

Besides support for the approach at the strategic, tactical and operational level, there must be vertical alignment: the respondents indicated that alignment and communication between those three levels within each organization are important. In addition, coordination is not only important to the network as a whole, but also within the involved organizations. An appointed internal coordinator within the organization can take a leading role when it comes to integrating the theme ‘healthy behavior’ into the daily work.

This internal coordinator serves multiple purposes, depending on the organization he or she works for. The common purpose of the internal coordinator is to bridge the gap between the tactical level of the approach (the program group) and the professional who work at the operational level. He or she supports professionals in paying attention to healthy lifestyle, providing concrete information on how to do this and inform colleagues about goals and ambitions of the approach.

According to the respondents, in the first three years of the implementation, the approach was seen as a program with financed projects. Subsequently, it is now seen just as a ‘movement, multiple organizations with a common goal’ (Professional at the tactical level #3). In addition to the network of organizations for connecting expertise, the approach is mainly embedded in the policies and regular processes of the organizations involved. In other words, health is considered in every activity or event, for example by offering healthy food and drinks.

However, the risk of embedding the approach in regular processes of the organizations is that attention for stimulating healthy behavior drops, because of more priority for other themes. For that reason, despite embedding, it is important to designate a coordinator in each of the organizations who is still responsible for ongoing attention among colleagues for the healthy behavior theme, according to the respondents.

Furthermore, respondents indicate that since the approach has been embedded among the involved organization, the local government no longer provides funding for activities but only for their core business, the ‘basic tasks’. Consequently, professionals experience being limited to spending fewer hours, resulting in a reduced impact on the mission of the Zwolle Healthy City approach. Professionals feel that with these additional tasks for the approach, they have to do more within the current agreements and funding.

All respondents agree that collaboration with private (local) organizations can be of added value to the approach. One respondent describes it as ‘a bonus, not a basis’. The bonus could be financial resources, knowledge, and more decisiveness, according to the respondents. Private organizations have a different working method than public organizations, i.e. they think more often in terms of action. At the same time, differences in culture, language and interests can be difficult when collaborating.

Within the respondent group, two different views exist on collaborating with private organizations. Some respondents state that it is important to ask private organizations with a certain expertise to pursue a specific goal in a project. Other respondents, on the other hand, believe that private organizations put up more resistance when their role is already fully specified, and it’s therefore better to let private organizations decide for themselves how they can contribute to the goals of the approach.

Many respondents mention the importance of collaboration with citizens in aligning the approach with citizens’ wishes and needs. So far, the Zwolle Healthy City approach has only done this in the organization and execution of specific interventions. This participation varies from drinking coffee with citizens to hear their wishes to starting a trajectory in which citizens come up with ideas which they implement and evaluate themselves. When organizing an event or running a project, the respondents consider it important to let citizens think along from the start and if possible throughout the whole process, i.e. in the design, organization, execution and evaluation. For example, with the Club2Move intervention, youngsters can exercise twice a week and participate in activities around healthy nutrition in a community sports club. The youngsters participate in promotion, organization and execution of the activities. Another example is when children and parents were invited to design, construct and maintain a natural playground in the neighborhood. It’s important to convert their input into action quickly, in the professionals’ experiences. Moreover, to make citizen participation as successful as possible, you have to involve citizens in small steps and, moreover, processes should not take too long.

One should not completely outsource the organization of activities to the citizens; a facilitating role of a professional is necessary, since citizens often want to contribute, but not fully organize and be responsible.

The respondents indicated that bringing the approach to people’s attention is especially important for raising awareness and commitment in the municipality and a sense of devotedness among the network organizations. One respondent said that an informative website about the approach contributes positively to profiling the approach like a brand, which is not the case in the municipality of Zwolle anymore. In the first three years of implementation, resources were available for external communication, which led to available hours for a communication advisor and a communication plan. By profiling the approach like a brand, success stories can be linked to the approach, which encourages other organizations to also want to be part of it.

On the other hand, some of the respondents indicate that it does not matter to the citizens whether an activity is part of the approach or not.

According to the respondents, the aim to align with "what is already there" is typical of the approach. The Zwolle Healthy City approach achieved this in three ways. Firstly, the approach was built up from a (smaller) network of organizations on a rather similar mission (but different scope), that was already there. As a result, some of the professionals involved already knew each other from previous collaborations. This made the start in 2010 slightly easier. Secondly, the approach tries to align with regular tasks, activities and working processes as much as possible. Thirdly, the local approach tries to link up with national developments. A national focus on a certain topic (e.g. the importance of drinking water or healthy school communities), can provide resources and impetus for the local implementation.

However, according to the respondents, you always have to consider what you need locally, because national programs do not always fit the local situation.

All respondents state that monitoring and evaluation of goals is important. They say it is important to show development in the objectives to keep involved organizations committed, and show them what they are contributing to. However, according to one respondent, the outcome effects that you measure (e.g. the percentage of children that is overweight) cannot simply be clarified by the activities and interventions undertaken “because there are all sorts of other factors that play a role” (Professional at the tactical level #3).

In addition to monitoring and evaluating the goals, monitoring and evaluating the process are just as important, according to the respondents. This allows the process to be continuously adjusted, newly derived topics to be put on the agenda and discussions about "where are we now and how do we move on" to provide insight and, therefore, a certain ‘grip’ on the approach. It is important to use these evaluations to improve activities and processes. If the evaluation shows that it is not working properly, the network of organizations should agree on doing things differently.

The fact that one of the partners of the Zwolle Healthy City approach is a research group from a university of applied sciences reinforces this key element. The researchers are responsible for monitoring the approach and students with various backgrounds are also involved. They do exploratory research on questions concerning a ‘healthy city’, raised by involved professionals or institutions, which stimulates efficient implementation.

Interviewing a varied group of professionals about eight years of implementation made it possible to identify perceived key elements for successful implementation of a community-based approach aimed at reducing socioeconomic health inequalities. Nonetheless, limitations of this study should be mentioned. Firstly, the design was retrospective, which means that the respondents had to retrieve experiences from their memory. This may have led to loss of details and muddled memory with the passage of time, since the first phase of implementation was more than eight years ago. Moreover, because the approach has been proven successful, respondents might mainly look back on the positive sides of the implementation and play down negative sides. To decrease this drawback, available documented conversations with (a selection of) the same professionals in 2013 were read by the interviewer (LW) beforehand, who could, therefore, help respondents to remember details of implementation in the first three years.

One strength of this study is that a considerable part of the professionals who contributed to the approach in 2010–2018 were invited for an interview and agreed to participate, and that they were employed by various organizations and on different occupation levels. This made it possible to formulate key elements from the perspective of the strategic, tactical and operational level. Including 29 respondents was sufficient because data saturation was reached after approximately 25 interviews. Another strength is that the respondents have been involved in the approach for a relatively long period of time (see Table 1), allowing them to give a good impression.

The overview of key elements described in this study may help and inspire other local integrated community-based approaches aimed at socioeconomic health inequalities or similar. Nowadays, more than one third of all Dutch municipality’s implement the JOGG approach, which the Zwolle Healthy City approach is largely based on. Professionals working in those municipalities can use the results in this study to improve the implementation process of this comprehensive approach. These results show which aspects on the process level are worth investing in, the key elements, and what factors contribute to strengthening those aspects. This applies not only to Dutch initiatives but also to other local integrated approaches worldwide.

Involved professionals working on the local Zwolle Healthy City approach can learn from the results as well, since the key elements provide guidance for improvement of the Zwolle Healthy City approach and accessory logic model. Further strengthening of the approach can take place based on the conclusions, and in collaboration with citizens, because our results indicate that citizens should be involved from the start and throughout the entire process, i.e. in the design, implementation and evaluation of activities.

The Zwolle Healthy City approach targets multiple sectors (e.g. public health, welfare, spatial planning, traffic and transport) and multiple levels of influence (interpersonal, organizational, community and policy). This type of local multi-level approach has long been the standard for tackling complex issues such as obesity and socioeconomic health inequalities. Recent literature has increasingly been paying attention to the relationships between those different levels in a whole-system approach [39–41]. In a whole-system approach, the central concern is to focus on the relationships between individuals and organizations working within a system [32]. The complexity of this type of approach lies not only in the relationships between the different levels, but also in how those relationships work in a certain context. Attwood et al. [42] state that we should not look at approaches as "one size fits all" but consider the issues within their wider environmental context. Questions as ‘how does it work’ and ‘in what context does it work’ are just as important as ‘what works’. For the evaluation of an approach in complex system thinking, this implicates that it is not only important to describe what the key elements of an approach are, but also which dynamic processes the impact of the key elements can reinforce or hinder (i.e. positive and negative feedback loops operating in the system).

For this reason, the main suggestion for further research is to investigate the mechanisms associated with the key elements and describe the context in which these mechanisms work.