https://orcid.org/0000-0002-7792-4741
Figures
Abstract
Introduction
Persons in socioeconomically disadvantaged situations are disproportionally exposed to obesogenic environments and often have fewer resources to counteract their influence, resulting in poorer dietary outcomes and lower levels of physical activity. This study aimed to map the complex systems within the municipal obesogenic environment driving unhealthy eating behaviors and insufficient walking among adults in socioeconomically disadvantaged situations in Flanders, and to identify potential actions for local policymakers.
Methods
Participatory group model building (GMB), a qualitative system dynamics method, was used to engage adults in socioeconomically disadvantaged situations (n = 7), local decision makers (n = 7) and local non-government stakeholders (n = 7) in two peri-urban Flemish municipalities. Two GMB sessions over two days were held in each municipality during which a causal loop diagram (CLD) was created and potential actions were identified for local policymakers by considering causal pathways and reinforcing loops in the system.
Results
The CLD represents participants’ perceptions of how the local obesogenic environment contributes to unhealthy eating and insufficient walking among adults in socioeconomically disadvantaged situations. Six interconnected subsystems were identified: 1) social context, 2) facilities and activities, 3) mobility options, 4) information, 5) affordability, and 6) marketing and promotions. The perceived obesogenic environment, adverse health effects, and disadvantaged socioeconomic context in which individuals live mutually reinforce each other. Social and physical isolation emerged as important cross-cutting mechanisms, interacting with financial constraints, limited mobility, inadequate facilities, non-inclusive activities and restricted access to information.
Conclusion
Participatory GMB and the resulting CLD provide a locally grounded basis for developing actions to promote healthy eating and walking among adults in socioeconomically disadvantaged situations. Actions addressing several subsystems simultaneously, while minimizing the effort required from residents to benefit from them, may be particularly valuable. Further co-design, feasibility assessment and sustained collaboration between residents, local stakeholders and decision-makers are needed to translate these findings into equitable local action.
Citation: D’Hooghe S, Inaç Y, Dury S, Deforche B, De Ridder K, Vandevijvere S, et al. (2026) Understanding the complex systems driving the perceived obesogenic environment among adults in socioeconomically disadvantaged situations: Findings from a participatory Group Model Building process in Flanders. PLoS One 21(10): e0346991. https://doi.org/10.1371/journal.pone.0346991
Editor: Elizabeth McGill, LSHTM: London School of Hygiene & Tropical Medicine, UNITED KINGDOM OF GREAT BRITAIN AND NORTHERN IRELAND
Received: March 25, 2026; Accepted: September 17, 2026; Published: October 5, 2026
Copyright: © 2026 D’Hooghe et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All generated and analyzed data during this study are included in this published article and its supplementary information files. Additionally, previous versions of the causal loop diagram can be accesses through Kumu.io (Group Model Building – CIVISANO project).
Funding: Sciensano internal grant.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Obesity, driven primarily by unhealthy eating and insufficient physical activity, is a complex, systemic and multi-causal issue [1]. Its prevalence worldwide is high and continues to increase, thereby contributing to an intensifying non-communicable diseases (NCD) epidemic [2]. In high-income countries, lower socioeconomic status is consistently associated with higher rates of overweight and obesity compared to a higher socioeconomic status, placing persons in socioeconomically disadvantaged situations at an elevated risk for NCDs and increasing health inequalities [3]. This pattern reflects both greater exposure to obesogenic environments and differential vulnerability, whereby fewer financial, time, mobility and psychosocial resources are available to buffer their influence [4].
It is increasingly recognized that both objective and perceived aspects of the physical activity and food environment influence eating behavior and physical activity [5–11]. Factors such as access to healthy foods, food prices, marketing, and social norms shape dietary behaviors, while walkability, transport infrastructure, availability and quality of recreational facilities, and perceived safety influence opportunities for physical activity [5,9,11–17]. Despite a substantial body of research describing the obesogenic environment and its dimensions, the prevention of obesity in community settings remains elusive [18]. Focusing on single environmental dimensions may miss the interactions between different factors influencing socioeconomic inequities in healthy eating behaviors and physical activity, and fails to incorporate dynamic elements, such as feedback and adaptations resulting from the interplay between people and their environments [19,20].
Systems thinking provides a valuable lens for studying such complexity. It allows researchers and stakeholders to conceptualize and visualize the dynamic relationships and feedback processes that reinforce obesogenic environments [21,22]. To date, most systems-based studies of obesogenic environments have been expert-led. A well-known example is the Foresight obesity map, developed for the UK government, in which experts mapped over one hundred interaction variables – clustered around a core energy-balance engine into domains such as food production, food consumption, individual psychology, physical activity and the built environment – to visualize why obesity is so resistant to single-focus interventions [23]. Other expert-led applications of systems thinking include systems map synthesized from literature reviews [10,19,24] and quantitative system dynamics and agent-based models [24–26]. These approaches have identified interacting macro-level drivers related to food production and marketing, the built environment, transport systems and socioeconomic conditions. Because such maps are primarily based on published evidence and expert knowledge, however, they may give less attention to how environments are perceived, navigated and experienced by the populations most affected.
Developing a system map with by people who know the system from within (e.g., local stakeholders, practitioners and community members) can situate the identified mechanisms within a specific decision-making context and promote shared understanding and ownership of both the model and its implications [27–29]. Studies comparing expert- and community-informed maps further suggest that participation influences the content of the resulting model. Community- and youth-led maps, for example, tend to place greater emphasis on social, environmental and psychosocial mechanisms, including stress, stigma and body image, whereas expert-developed maps give more prominence to physiological and individual-level determinants of the same problem [30–32].
Group Model Building (GMB) has emerged as a valuable participatory method within systems thinking, where a diversity of stakeholders participate in collaboratively conceptualizing causal loop diagrams (CLDs) to describe complex problems, generate dynamic hypotheses, and identify potential solutions [33]. GMB has its origins in operations research for optimizing industrial processes, but it is increasingly applied within health settings over the past decade. Studies have explored complex systems driving childhood obesity [22,28,34–36], unhealthy eating behavior [19,37,38] and insufficient physical activity [39]. These studies typically identify interrelated subsystems such as housing and built environment [12], transport infrastructure [12,26], health literacy [23], employment and work conditions [12,26], marketing and food promotion [23], and social norms [23,26]. Some of these studies extend beyond descriptive mapping by extracting key actions [28,37,39] or intervention development ideas [35].
Nevertheless, relatively few GMB studies have directly involved adults in socioeconomically disadvantaged situations in mapping the obesogenic environments that shape their everyday lives [40]. Consequently, limited evidence is available on how their perspectives influence the structure of system maps or the actions identified for them [41]. This restricts the development of community-based interventions that are both contextually appropriate and equity oriented [42,43].
Evidence is also limited for peri-urban settings, which combine urban and rural characteristics and remain underrepresented compared with urban environments. Such areas are becoming increasingly common across Europe and may face specific structural and mobility-related challenges, including limited transport, dispersed services and poorly accessible infrastructure [44–47].
The aims of this study were therefore (1) to jointly conceptualize the municipal obesogenic environment driving unhealthy eating and insufficient walking among adults in socioeconomically disadvantaged situations in two peri-urban Flemish municipalities through GMB, and (2) to explore actionable points for local governments to reverse current trends. Although walking, whether recreational or for transport, represents only one dimension of physical activity, it is an important component of everyday movement and strongly shaped by local environmental conditions [34].
The present study was conducted within the CIVISANO-project, a broader project that used mixed-methods to examine the role of objective and perceived environmental factors in physical activity and eating behavior among adults in socioeconomically disadvantaged situations in two peri-urban Flemish municipalities, Duffel and Herselt. The study protocol is described in detail elsewhere [48]. CIVISANO was conducted in close collaboration with local policymakers, stakeholders and residents, reflecting a shared concern about unhealthy eating and insufficient physical activity in the respective municipalities and the need for solutions targeting environmental determinants. Previous CIVISANO findings explored perceptions of adults in socioeconomically disadvantaged situations towards the local environment and its role in recreational walking behavior [49], walking for transport [50] and eating behavior [51]. GMB was used in the final phase of the project to integrate these findings with the perspectives of residents and municipal stakeholders and to identify locally relevant opportunities for action.
Materials and methods
Settings
The study took place in Duffel and Herselt, two medium-sized peri-urban municipalities in the province of Antwerp in the Flemish region of Belgium.
Duffel is a monocentric municipality geographically divided by a river, with most services, shops, and public facilities concentrated near the town center on the east side. It has a higher population density (796 inhabitants/km²) and includes several supermarkets and a train station, which provide relatively good access to services. However, green space is more limited and mainly concentrated along the river. Social housing is located both in and outside the center, making car or public transport use important for daily mobility. Duffel also has a large psychiatric hospital, and many former patients remain in the municipality through supported housing arrangements.
Herselt, in contrast, is more dispersed and consists of several sub-municipalities surrounded by wooded areas and agricultural land. It has a lower population density (283 inhabitants/km²), extensive green space, and multiple recreational outdoor areas, but fewer centrally located shops, fewer supermarkets, and no train station. Distances between facilities are longer, and public transport connections between sub-municipalities are limited. In addition to social housing, a notable feature of Herselt is the presence of low-income households living in campers that were originally intended as second residences. These housing situations are not fully regulated and often fall outside standard municipal housing frameworks.
Socioeconomic conditions also differ. Based on the deprivation index (percentage of children born in deprived households), Duffel (13.6%) is slightly above the Flemish average (≈12%), while Herselt (7.5%) is below it. Across both municipalities, adults in socioeconomically disadvantaged situations were mostly concentrated in social housing sites or dispersed rental apartments, often located outside central service areas, increasing reliance on mobility options to access food, activities, and public services.
Ethical approval
This study was approved by the Medical Ethics Committee of Ghent University Hospital (BC-09260) and was conducted in line with the recommendations of the Belgian Data Protection Authority. All participants signed a written informed consent form prior to participation.
Participant selection and recruitment
Local residents in socioeconomically disadvantaged situations and stakeholders from the local government and local non-government organizations were invited to participate during the period from June to August 2023. We critically reflected beforehand on who to invite for the group model building sessions, as we wanted to ensure an open and respectful space to exchange ideas and perceptions. Therefore, we invited adults in socioeconomically disadvantaged situations who previously participated in focus groups within the CIVISANO-project [49,52] and showed interest to exchange and discuss with local stakeholders. Participants in these earlier CIVISANO focus groups had been recruited through local intermediary organizations (e.g., social welfare services, poverty organizations and community centers), municipal channels, and snowball sampling, using the same socioeconomic eligibility criteria as described below [49,52]. For the GMB sessions, focus group participants who had expressed such interest were re-contacted by the research team. This prior participation familiarized them with the topics and with the experiences of other participants. In addition, previous findings were shared with stakeholders and policymakers involved in this study, ensuring that participants were informed about earlier results and pre-identified challenges faced by people in socioeconomically disadvantaged situations. The participants in socioeconomically disadvantaged situations had to be between 25 and 65 years old, reside in one of the involved municipalities and be considered socioeconomically disadvantaged by meeting at least two of the following criteria: no higher educational degree [53], no current paid or very low paid work [54], current perceived financial difficulties [55] and a perceived socioeconomic status of 5 or less on a 10-point scale (with 10 indicating the highest perceived SES and 1 indicating the lowest) [56]. In total, 4 adults in socioeconomically disadvantaged situations in Duffel and 3 adults in socioeconomically disadvantaged situations in Herselt agreed to participate. Among these 7 participants, there were four women (ages 31, 51, 56, and 63) and three men (ages 46, 56, and 62). In Duffel, 4 policy makers and 2 stakeholders from a local organization participated, and 3 policy makers and 5 stakeholders from local organizations participated in Herselt. In total, there were 10 participants in Duffel and 11 participants in Herselt, which corresponds with the recommendation to have at least 6 and at most 15 participants to conduct a group model building workshop [57]. The participant characteristics are shown in Table 1.
Group model building workshops
Workshop design and facilitation.
One two-session Group Model Building workshop was conducted in each municipality in August 2023. Each session lasted approximately three hours. The workshops aimed to develop a shared representation of the drivers and consequences of unhealthy eating and insufficient walking within the perceived municipal obesogenic environment and to generate potential actions for local policy makers.
The sessions were designed, organized and facilitated by the core modelling team, comprising the lead researcher (SD), a supporting researcher (YI) and an expert in GMB (LG).
The workshop design was informed by the Group Model Building principles and facilitation strategies outlined by Vennix et al. (1994), Andersen and Richardson (1997) and Luna-Reyes et al. (2006). Predefined but adaptable scripts were used to structure the sessions. All scripts can be accessed online from Scriptapedia (https://en.wikibooks.org/wiki/Scriptapedia). The sequence of workshop activities and scripts used are presented in Table 2.
Problem framing and system boundary.
The first session began with an introduction to the objectives of the GMB sessions and the process of modeling. Findings from previous components of the CIVISANO-project were briefly presented to provide a shared starting point for the discussion.
Additionally, a ‘hopes & fears’ exercise took place in which participants could share expectations and express hopes and concerns regarding the GMB process and its potential outcomes (S1 Fig). In both workshops, participants were presented with a reference mode, as a simple graphical representation of the core problem over time (S2 Fig). This exercise supported the development of a shared understanding of the problem before variables and causal relationships were elicited.
The system boundary used for the GMB sessions was defined as the municipal environment, as perceived and experienced by adults in socioeconomically disadvantaged situations. Unhealthy eating and insufficient walking were defined as the two focal problem behaviors. Individual-level determinants and national or supranational policy levels were outside the primary scope of the exercise. However, participants could, and did, raise them when they considered them relevant because of how they operated through their local environment.
No predefined definition of “healthy” or “unhealthy” eating or walking was imposed. Participants were instead asked to discuss these behaviors according to their own understandings and experiences of practices that they believed supported or undermined health.
Variable elicitation and development of the municipal CLDs.
Once the problem was described and agreed on, participants were led in a variable elicitation exercise using the nominal group technique where they were asked to identify the main drivers and consequences within the perceived obesogenic environment of respectively unhealthy eating and insufficient walking in their municipalities (S3 Fig). Participants were encouraged to think in terms of causes and consequences and to explain how identified factors were connected.
The variables generated during the exercise were used to construct a causal loop diagram for each municipality. Variables were added to a large poster at the front of the room, after which participants identified causal directions and connections between them. The moderators supported the diagramming process by asking clarifying questions, helping participants articulate the direction of relationships and organizing the emerging structures visually. With support from the moderator, the variables and relationships were organized into causal pathways and feedback loops (S4 Fig).
Feedback loops operate between system variables. In reinforcing loops, a change in a variable propagates through the loop to amplify the initial change in the same direction, producing growth or decline; in balancing loops, the change propagates to counteract the initial change, as such an increase feeds back as a decrease and vice versa, giving the loop a stabilizing, goal-seeking character [43].
Digitization and participant review of the municipal CLDs.
After the first session, SD transferred the hand-drawn causal loop diagrams to a digital format using Kumu.io (2023). The moderator team used notes from the sessions to identify and name preliminary subsystems and feedback loops. The discussions were also audio-recorded, which allowed the research team to clarify and confirm notes where necessary.
During the second workshop, the digitized CLDs and the changes proposed by the moderator team were presented to participants for collective review (S5 Fig). Participants in both municipalities suggested edits and added some new variables (S6 Fig). The resulting municipal CLDs therefore reflected an iterative process involving participant elicitation, moderator-supported organization and subsequent participant review.
Generation of action ideas.
Throughout the process, the CLDs were treated as shared dynamic hypotheses about the factors and processes contributing to unhealthy eating and insufficient walking in the municipal environment. They were not considered empirically validated causal models. Identifying leverage points through a purely qualitative exercise is challenging and may lead to overinterpretation when a model has not undergone subsequent quantitative modelling or empirical testing.
Instead, participants used the causal pathways and feedback loops represented in the municipal CLDs as a basis for generating and prioritizing possible actions. This process followed a structured nominal group exercise based on the Scriptapedia “Action ideas” script. The exercise was intended to move from problem representation towards potential local responses while retaining a clear connection with the factors and relationships represented in the CLD. The resulting actions should therefore be interpreted as participant-generated, exploratory intervention ideas rather than as empirically validated leverage points.
Workshop evaluation.
At the end of the GMB sessions, all participants were invited to complete a brief, voluntary, and anonymous evaluation survey. Also, a short group discussion took place to revisit the hopes and fears identified at the beginning of the process and to reflect on whether participants’ initial expectations and concerns had been addressed.
Development of the merged summary CLD.
After the sessions in both municipalities, the modeling team integrated the final municipal CLDs into a single summary CLD. The purpose of this integration was not to compare Duffel and Herselt, but to synthesize the similarities between the two local models and create a more comprehensive representation of the shared system.
To maintain readability, the summary CLD only includes the causal connections required to represent the identified subsystems and named feedback loops. This approach is consistent with Finegood et al. (2010), who stated that by reducing the complexity of the visual image of the CLD, while preserving its underlying information, key relationships become more apparent [58].
The merged CLD was subsequently shared with participants for review, with the aim of determining whether it remained recognizable as a representation of the workshop discussions. No comments or requests for changes were received.
Rigor and reflexivity
Rigor and traceability were supported through established Group Model Building practices, including predefined and adaptable Scriptapedia scripts, audio-recoding of the sessions, verification of the notes against recordings, documentation of the successive model versions (S5 and S6 Appendices in S5 and S6 Figs and the version history on Kumu.io), and participant review of the moderator-edited municipal CLDs and the merged CLD.
During manuscript revision, variable names and link polarities were also edited to improve consistency with causal loop diagramming guidelines. Specifically, directional wording was removed from variable names and corresponding polarity adjustments were made. Additionally, delay marks were added to relationships for which the modelling team considered that the effect would plausibly develop over time. These delays were identified through discussion within the research team and were not explicitly elicited or subsequently reviewed by participants; they should therefore be interpreted as modeler-added annotations. These editorial changes did not alter the causal structure previously reviewed by participants.
The core modelling team had established relationships with participants through previous CIVISANO focus groups. None lived in the participating municipalities, and their educational and socioeconomic positions differed from those of the participating residents. The team recognized that this positional distance, together with its environmental framing of obesity, could influence elicitation and interpretation. Participants’ own terminology was therefore retained where possible, no definitions of healthy or unhealthy behavior were imposed, and moderator interpretations were returned to participants for review. As in any qualitative inquiry, however, the subsequent processing and analysis of the workshop data — including the naming of subsystems and feedback loops and the synthesis of the two municipal CLDs — constituted necessary interpretive work by the research team, anchored in participants’ shared understanding through member checking and validation of the merged model.
Results
Participants discussed and agreed that obesity-related behaviors, in particular unhealthy eating behavior and insufficient walking, have increased over time. Additionally, they agreed that the food and physical activity environments have become unhealthier over time, although they described this as more pronounced within the food environment. At the end of the GMB sessions, participants in Duffel produced two causal loop diagrams respectively one for the role of the perceived obesogenic environment in healthy eating behavior and one for recreational walking, while participants in Herselt chose to elaborate one causal loop diagram incorporating both behaviors (Appendix 5 and 6 in S5 and S6 Figs). The final summary CLD where all the overarching subsystems and feedback were combined is focusing on the perceived obesogenic environment that leads to both unhealthy eating behavior and a lack of walking presented in Fig 1. To enhance the clarity and readability of the diagram, not all connections between variables have been depicted. The full map is available online at Kumu.io.
All labelled loops are described in full in Table 3.
Among our sample of participants in socioeconomically disadvantaged situations, the majority experienced health issues that exacerbated their socioeconomic disadvantage and financial hardship. Both the socioeconomically disadvantaged situation and the health problems negatively affected how they used and perceived their local environment. This environment was often perceived as obesogenic, contributing to unhealthy eating behaviors and reduced walking, which in turn worsened their health outcomes.
The following subsystems were identified by the researchers and agreed on by the participants: subsystem 1 – social context, subsystem 2 – facilities and activities, subsystem 3 – mobility options, subsystem 4 – information, subsystem 5 – affordability and subsystem 6 – marketing and promotions. Reinforcing loops (i.e., if a variable increases, the effect through the cycle will return and increase the same variable) are labeled with an ‘R’ and Balancing loops (i.e., if a variable increases, the effect through the cycle will return and decrease the same variable) with a ‘B’ [59]. An overview of all feedback loops identified in the CLD is provided in Table 3, including each loop’s type, the subsystems it spans, its full pathway with link polarities and delays, and its interpretation. In the subsystem descriptions below, loops are referenced by these labels (R1–R8, B1–B2).
A simplified version with the identified subsystems and how they interact is presented in Fig 2. This CLD reflects the joint discussion and perception of participants regarding how different elements of their environment contribute to an obesogenic environment. According to the participants, the obesogenic environment is influenced by, and in turn influences, all subsystems directly. Participants also described interconnections between subsystems. For example, they perceived the social context as being directly shaped by adverse health effects, a lack of transport options, low affordability, and a lack of adequate and accessible facilities and activities.
Participants generated potential actions in groups using nominal group technique, informed by the causal pathways and feedback loops in the CLD, and exchanged ideas afterwards. The action ideas presented below (Tables 4–9) are participant-generated, exploratory intervention ideas. The level of change was determined by the aimed objective of the proposed action, which focused on either individual behavior change or community-level change (including physical, economic, sociocultural, and informational aspects). This categorization aligns with previous studies conducted in the context of the CIVISANO project, which utilized the Analysis Grid for Environments Linked to Obesity (ANGELO framework). Consequently, most participants were already familiar with these categories.
Subsystem 1 – Social context
During the workshops, participants explained that poor health and related problems (such as chronic conditions, pain, or fatigue) are linked to mobility issues and a greater dependence on other people. These aspects were described as contributing to feelings of lower self-esteem and shame. According to the participants, such feelings discourage social interactions and therefore increase social isolation. Several participants emphasized that this social isolation is not only problematic in itself, but also because it makes asking for help more difficult. As they explained, being isolated and dependent on others can increase the threshold to reach out, whether for practical support or to access public centers such as social welfare services or local food distributions. This, in turn, reinforced the negative perception that public spaces and facilities are inaccessible or not inclusive, which further discourages walking and healthy eating. Some participants also shared that a lower self-esteem and shame often directly trigger unhealthy eating behavior, such as eating for comfort or to reduce stress. At the same time, participants stated that small, everyday positive social encounters (for example greeting neighbors or chatting with passersby) leads to an increased motivation to get outside and walk, reducing the social isolation and ameliorating the perception of accessibility and inclusiveness of public spaces and facilities (e.g., parks or public open spaces) (R1). They also highlighted that these spaces could serve as meeting spots, further promoting such positive interactions (R1). Participants stressed that these positive encounters increase their feelings of safety, while negative social interactions (such as experiencing stigma or discrimination) or public drug and alcohol abuse evoke fear or feelings of threat. These negative experiences reinforced their perception of public spaces as unsafe, which discouraged walking (R2). Conversely, a perceived sense of safety will also reinforce the positive perception of the social environment, as more encounters can take place in a safe manner (R2). Lastly, participants stated that watching people’s unhealthy eating habits in their surroundings and observing more and more unhealthy food around them seem to normalize this. This, in their perception, created a kind of social acceptance, which in turn reinforced unhealthy eating behaviors, the same reinforcing dynamic the participants linked to industry profits (R4).
Areas for action.
Stimulating social interactions, fighting against social isolation, and increasing the feelings of physical and social safety were identified by the participants as potential interventions points to promote walking and healthy eating behavior (Table 4).
Subsystem 2 – Facilities and activities
According to the participants, persons with health and mobility issues have an increased need for adequate and safe facilities, such as the sufficient presence of sanitary facilities or benches to rest. They described how the fear that these facilities are not present or within reach contributes to people walking less (R3). Additionally, participants also perceived that persons with health and mobility issues are more likely to need activities that are adapted to their conditions (e.g., a walking group that organizes activities with varied speeds and distance). The perceived inappropriateness of facilities and activities was mentioned by participants in socioeconomically disadvantaged situations as a barrier because of feelings of shame about one’s condition, which participants felt reinforced physical and social isolation, feeding back into the social-context dynamics, described in Subsystem 1 (R1, R2).
In combination with limited mobility options (e.g., no car or bicycle ownership or insufficient public transportation), participants expressed that these factors contribute not only to a perceived lack of accessibility of public spaces and facilities, but also to actual structural constraints that limit where people go and what they can access. This in turn discourages walking and supports unhealthy eating.
Areas for action.
The creation of adequate and inclusive activities, with the provision of rest areas and sanitary facilities, was identified as intervention points by the participants (Table 5).
Subsystem 3 – Mobility options
Participants stated that people who experience mobility issues may perceive distances as too long to cover on foot. They described how this perception contributes to physical and social isolation, lowering the perceived accessibility of (healthy) food stores, facilities, or public spaces, and thereby reinforcing the perceived obesogenic environment and therefore unhealthy eating and insufficient walking. This chain constitutes a part of the reinforcing loop R8, described in full in Subsystem 5.
At the same time, participants emphasized that this is not only a matter of perception. They pointed out concrete structural barriers, such as limited car and bike ownership due to low affordability and inadequacies in public transport in terms of the areas served and the time schedule. Additionally, participants explained that their dependence on limited public transport offerings further reinforced physical and social isolation. These limited mobility options were described as reducing opportunities to use diverse facilities (such as stores, local community centers and parks) that are not in their immediate vicinity, which participants felt further reinforced insufficient walking and unhealthy eating.
Areas for action.
To improve the mobility options and perceived proximity was identified as an intervention point for this subsystem by the participants (Table 6).
Subsystem 4 – Information environment
Participants stated that a lack of awareness of walking trails or attractive walking facilities or public places in the neighborhood contributes to insufficient walking. They pointed out that this is not merely a matter of residents’ awareness, but also pertains to the availability and accessibility of information within the municipality regarding existing facilities and walking activities. Additionally, they mentioned that insufficient walking will also contribute to this lack of awareness, as people won’t go out and therefore do not ‘discover’ their neighborhood. Healthy eating behavior also involves the availability and accessibility of information on healthy foods, where it can be found and how it can be used. According to our participants this will promote healthy eating behavior. Participants assumed that persons with health problems who seek and receive medical help from a doctor or dietician have access to this information and can thus increase their knowledge and change their health behavior (B1). A related, larger balancing loop operates across several subsystems (B2). As unhealthy eating and insufficient walking increase, health problems accumulate (with delay), leading more residents to use medical care and thereby gain access to information and knowledge. This improved knowledge reduces the perceived walking distance to destinations, which lowers physical and social isolation, improves the perceived accessibility of public spaces and the perceived availability of facilities, and ultimately counteracts the initial rise in unhealthy eating and insufficient walking. Furthermore, it was also cited that while community members know the major supermarkets in the municipality, there is very little awareness of and familiarity with (healthy) local alternatives (e.g., local farmers, community gardens or picking fields), which also reinforces unhealthy eating.
Areas for action.
Providing accessible information and disseminating and communicating about this in an effective way were identified as key intervention points within this subsystem (Table 7).
Subsystem 5 – Affordability
Participants in socioeconomically disadvantaged situations emphasized that a disadvantaged socioeconomic situation of community members directly reduced both their actual and perceived ability to afford healthy options. They explained that this lack of ability to afford healthy options contributed to unhealthy behaviors, such as frequent consumption of unhealthy food or insufficient physical activity, which were perceived to contribute to poor health. In turn, poor health was described as making it harder to find or maintain employment, which reinforced a disadvantaged socioeconomic situation, creating a reinforcing feedback loop (R7) in which lower disposable income, reduced affordability of healthy options, unhealthy behaviors and health problems perpetuate one another over time. The health and employment effects in this loop operate with considerable delay.
They noted that promotions and discounts were more common for unhealthy food. In addition, participants explained that people with specific dietary needs (e.g., sugar-free, lactose-free or vegan products) often had to purchase more expensive alternatives, which they felt contributed to the perception that healthy eating is less affordable.
Costs related to transportation were also highlighted. Participants perceived that travel costs, such as fuel or mileage costs for cars and the price of public transportation, limited their access to mobility. Tracing this link further, reveals that it is embedded in a larger reinforcing loop (R8). Higher travel costs increased perceived walking distances, creating barriers to reaching facilities outside their immediate environment. As a result, they felt discouraged from going out, which reinforced physical and social isolation and limited opportunities for walking and healthy eating. Over time, the resulting health problems reduce disposable household income, which in turn makes travel costs weigh more heavily.
Additionally, the participants linked a limited budget and a lack of disposable income to living in social housing with often unfavorable housing conditions such as restricted kitchen space (e.g., limited storage space or lack of a freezer) and limited cooking equipment. They explained that this made home-cooking more difficult and reduced the affordability of buying larger or frozen quantities, which were perceived to be cheaper.
Finally, participants described how financial constraints also influenced ‘time and mental affordability’. They explained that work or household responsibilities took priority, and there was less energy or time available for walking or home-cooking. Moreover, they stated that a limited budget makes it more difficult to buy appropriate walking equipment (e.g., hiking shoes or rain equipment), which they felt reduced the ability to engage in walking. Additionally, participants mentioned that the presence of unhealthy prepared products (such as fast food or ready-to-eat meals in stores) in combination with limited time, other priorities and insufficient cooking equipment contributed to unhealthy eating behavior.
Areas for action.
The provision of affordable equipment and alternatives for healthy eating behavior were identified as important intervention points (Table 8).
Subsystem 6 – Marketing and promotion
Participants mainly referred to the role of the food industry and advertising in relation to eating behaviors. They explained that an increasingly obesogenic environment, in which unhealthy food is omnipresent and visibly consumed, raises the social acceptability of unhealthy eating, which in turn reinforces unhealthy eating and insufficient walking. Participants believed that this ultimately benefits the food industry through increased profits, and that these profits in turn stimulate the industry to further shape the obesogenic environment, closing a reinforcing feedback loop (R4). In addition, according to the participants, the industry reinvests these profits in marketing strategies, such as offering more discounts on unhealthy food and promoting larger package sizes. They felt that these practices make unhealthy food appear more affordable compared to healthier options and further increase the social acceptability of unhealthy eating, encouraging overconsumption (R5). Participants also expressed concerns that such marketing strategies can mislead consumers. For example, they described how promotions may give you the impression that products are cheaper, even when this is not the case. In their views, these tactics contribute to misinformation that undermines people’s knowledge about healthy eating. They explained that this misinformation can counteract the positive influence of accessible information in the environment, ultimately reinforcing unhealthy eating behaviors (R6).
Areas for action.
Notably, all actions proposed within this subsystem aim to make healthy food more attractive or more visible (Table 9). No participant proposed limiting marketing practices they had themselves described as harmful.
Discussion
This study used Group Model Building to explore adults in socioeconomically disadvantaged situations, local policymakers and community stakeholders perceived the municipal systems contributing to unhealthy eating and insufficient walking in two peri-urban Flemish municipalities. Six closely connected subsystems were identified: 1) social context, 2) facilities and activities, 3) mobility options, 4) affordability, 5) information, and 6) marketing and promotion. The main contribution of the study lies not in identifying these determinants individually, many of which are already well documented, but in showing how they interact and accumulate in everyday life. In particular, limited mobility, financial constraints, restricted access to information and social isolation were perceived to reinforce one another, reducing residents’ opportunities to access healthy food, participate in activities and maintain social connections.
Recurrent and context-specific system dynamics
Several of the identified mechanisms are consistent with previous systems maps of obesity, dietary behavior and physical activity. These include reinforcing relationships between socioeconomic disadvantage, stress and poor health; the interaction between food affordability, exposure to unhealthy food and dietary preferences; the influence of social norms and marketing; and the way transport and the built environment shape access to food and opportunities for physical activity [10,60–63]. An expert-developed map of obesity in socioeconomically disadvantaged urban neighborhoods in the Netherlands similarly identified mutually reinforcing dynamics across the food, physical activity, socioeconomic and sociopolitical environments. It highlighted, among other mechanisms, the interaction between unhealthy food supply and demand, chronic stress, income insecurity, limited facilities and institutional distance [62]. A recent participatory study among people living on a low income also identified scarcity and wellbeing, the unhealthy food environment, and social and societal barriers as three central and interacting dynamics shaping dietary behavior [64].
The similarity between the municipal CLDs also suggests that some dynamics extend beyond the specific characteristics of Duffel and Herselt. This is consistent with Calancie et al., who compared participatory CLDs developed in seven communities across the United States and identified recurring dynamics related to intergenerational poverty, difficulty meeting basic needs, stress and mental health, healthy food supply and demand, safety, social norms and community engagement [35]. Such convergence does not imply that local context is unimportant. Rather, it suggests that common mechanisms may be expressed differently depending on the spatial, institutional and socioeconomic characteristics of a community.
Within the present study, mobility appeared to be a particularly important cross-cutting mechanism. In the peri-urban municipalities studied, dispersed settlement patterns, longer distances, limited public transport and the location of some social housing outside central service areas affected more than transport alone. Mobility shaped access to affordable food, recreational opportunities, public services, information and social networks. It therefore functioned as a gateway connecting several subsystems. These findings extend the concept of mobility poverty by illustrating how transport disadvantage can compound other forms of socioeconomic disadvantage and contribute simultaneously to unhealthy eating, insufficient walking and social isolation [47,65].
The centrality of social isolation is another notable feature of the resulting model. Previous studies have often emphasized the protective role of social support, family modelling and community organizations [13,34]. Participants in this study instead described how isolation could emerge from the interaction of inaccessible transport, physically dispersed housing, limited information, financial constraints, stigma and a shortage of inclusive opportunities for social participation. Social isolation was therefore not represented solely as an interpersonal problem but as an outcome of the organization of the wider local environment. It also fed back into other subsystems by reducing opportunities to exchange information, participate in activities, seek support and build the confidence needed to use available services.
Implications for action across subsystems
From a systems perspective, actions addressing several subsystems simultaneously may offer greater potential than isolated interventions. Four participant-generated actions were particularly cross-cutting: a multifunctional service center offering tailored activities; a mobile service combining an information point with access to affordable food across the sub-municipalities; clearly signposted walking routes with information on distances, benches and sanitary facilities; and home delivery of affordable healthy meals combined with proactive outreach.
These actions could generate multiple, mutually reinforcing benefits. For example, a mobile food and information service could simultaneously reduce transport barriers, improve access to affordable food, provide information and create opportunities for social contact. Similarly, a multifunctional service center could combine practical support, inclusive activities and opportunities to strengthen social networks. These actions should not yet be interpreted as formally identified leverage points, as the study did not assess their likely effect sizes, feasibility, costs or unintended consequences. They nevertheless represent promising candidates for further co-design and feasibility testing because they respond to several of the barriers represented in the CLD rather than addressing one factor in isolation.
The findings also illustrate the relevance of proportional universalism. Community-level improvements may benefit the population generally, but additional support is likely to be necessary for residents experiencing overlapping financial, mobility, health and social barriers [66]. The Depth Framework further distinguishes interventions according to their exposure, mechanism and level of engagement [67]. Many actions proposed in the workshops would improve the environment but would still require residents to identify, understand and actively use the available support. Their effectiveness may consequently remain dependent on individual time, confidence, health, mobility and cognitive capacity.
To reduce these agentic demands, environmental improvements could be combined with proactive and tailored forms of outreach. For example, providing information about existing activities may have limited effect when residents are socially isolated, lack transport or are uncertain whether they will feel welcome. Combining accessible activities with transport support, personal outreach and affordable participation could address these barriers together. Similarly, increasing the availability of healthy food is unlikely to be sufficient when affordability, transport and opportunities for food preparation remain constrained. Coordinated actions across policy domains are therefore needed rather than a collection of disconnected initiatives.
Interpretation of the feedback structure
Participants initially identified one balancing loop, in which people experiencing health problems could obtain professional guidance, increase their knowledge and make changes that support healthier behavior. One additional balancing dynamics were identified when the diagram was reviewed during manuscript revision. The predominance of reinforcing loops may have several, non-exclusive explanations. First, it may represent the underlying system structure: a local environment that largely promotes unhealthy eating behaviors and insufficient walking, and therefore requires deliberate action to shift existing dynamics. Second, the design and focus of the elicitation process may have influenced the types of feedback dynamics identified. The sessions were primarily oriented towards eliciting participants’ perceptions and causal narratives, rather than systematically probing for system limits or balancing mechanisms. Third, modelling and naming decisions can influence the apparent polarity of a loop. Directionally phrased variables may lead a balancing relationship to be represented as a sequence of positive causal links within an apparently reinforcing narrative. Finally, the small number of balancing loops should therefore not be interpreted as evidence that no stabilizing mechanisms were present in the local system. The model remains a simplification representation of perceived causal relationships, based on participants’ perceptions and may therefore omit relevant mechanisms or represent them imperfectly.
Limited representation of political and environmental change
No variables explicitly represented local political decision-making, policy investment or institutional accountability, despite the participation of policymakers in the sessions. Moreover, many proposed actions focused on information, awareness or individual support rather than on changing upstream environmental conditions. Similar tendencies have been observed in other participatory studies [28,68]. One explanation for this tendency could be that individuals, whether residents or policymakers, are accustomed to framing obesity in terms of personal responsibility, which is a dominant narrative in public health discourse [69,70]. Previous research shows that systemic and upstream determinants are often less visible and more difficult for laypeople and local stakeholders to conceptualize compared to concrete, individual-level solutions [71]. For local policymakers, interventions targeting individual behaviors are often seen as more straightforward, requiring fewer resources and shorter implementation timelines, whereas systemic changes demand significant infrastructure investment and longer-term commitment [72]. This may also be reinforced by limited budgets and rigid decision-making structures, as they themselves expressed during the ‘hopes and fears’ discussion. In this context, proposing individual-level actions may have felt more feasible and realistic than suggesting broader environmental interventions that require higher-level resources or structural policy changes. A complementary explanation lies in the composition of the sessions themselves: although facilitation measures were designed to mitigate power imbalances and residents reported feeling heard, proposing environment-changing actions often implies critique of existing policy choices, which may have felt less comfortable for residents with policymakers in the room — and, conversely, policymakers may have been reluctant to propose actions committing resources in front of residents. Separate preparatory sessions per participant group, followed by a joint session, could be considered in future GMB processes.
Strengths and limitations
A major strength of the study was the direct involvement of adults in socioeconomically disadvantaged situations alongside local policymakers and community stakeholders. This combined lived knowledge of the local environment with knowledge of existing services, institutional constraints and municipal decision-making. Bringing the groups together also created a dynamic of mutual learning where residents gained insight into the institutional considerations behind local decisions, while policymakers were confronted with how their municipality is experienced in daily life. The focus on two peri-urban municipalities also contributes evidence from settings that remain less studied than densely urban environments. Conducting two sessions in each municipality allowed participants to review and revise the municipal CLDs and to return to the process after a period of reflection.
Several limitations should be considered. First, the participant groups were small and purposively selected. Participants in socioeconomically disadvantaged situations had previously participated in CIVISANO research and had expressed an interest in engaging with policymakers. They may therefore have had more confidence, time or social support than other adults experiencing socioeconomic disadvantage. Residents were also numerically outnumbered by professional and institutional participants, and mixed-group power dynamics may have affected what was discussed despite the inclusive facilitation measures used. Related to this, participants in socioeconomically disadvantaged situations contributed primarily from lived experience. Facilitation supported the step from personal experience to general system structure (variables were phrased in generic terms, moderators probed whether mechanisms applied to residents in similar situations), and most participants made this abstraction readily, frequently generalizing to peers and neighbors. Nevertheless, the CLD reflects the range of experiences present in the room, and mechanisms affecting groups not represented may be underrepresented.
Second, the time available limited how deeply participants could explore feedback structures, prioritize actions and assess their feasibility or possible unintended effects. The study did not undertake a formal leverage-point analysis, and the proposed actions should be interpreted as exploratory ideas. Embedding GMB earlier and more iteratively within a research or policy process could create more opportunities to refine, implement and monitor actions.
Third, the identification and naming of subsystems and feedback loops, and the reconciliation of the two municipal models into a merged CLD was produced through an interpretive synthesis by the modelling team. Although it was shared with participants and no comments or requests for changes were received, the absence of feedback should not be interpreted as formal validation.
Finally, a qualitative CLD represents perceived causal structures but does not estimate the strength of relationships, delays, loop dominance or the net effects of interacting feedback processes. A subsequent quantitative system-dynamics model, informed by local data, could examine how these dynamics evolve over time and test how alternative policy combinations might affect the system [73].
Conclusion
This study shows how adults in socioeconomically disadvantaged situations and local stakeholders perceive unhealthy eating and insufficient walking as outcomes of interacting social, spatial, financial and informational conditions. Six interacting subsystems were identified: 1) social context, 2) facilities and activities, 3) mobility options, 4) affordability, 5) information and 6) marketing and promotion.
Many of the identified mechanisms resemble those found in other obesity-related systems maps, supporting the presence of recurring dynamics across settings. The study adds to this evidence by showing how peri-urban spatial dispersion, limited mobility and social isolation can operate as cross-cutting mechanisms that restrict access not only to food and physical activity opportunities but also to information, services and social connection.
The findings suggest that local action should move beyond isolated information or behavior-change initiatives. Actions that reduce financial, mobility, informational and social barriers simultaneously, and that minimize the effort required from residents to benefit from them, may be particularly valuable. These actions require further co-design, feasibility assessment and evaluation, but they provide a locally grounded basis for developing more integrated and equitable responses.
Supporting information
S1 Fig. De rol van de voedingsomgeving in Duffel.
https://doi.org/10.1371/journal.pone.0346991.s001
(PDF)
S2 Fig. De rol van de wandelomgeving in Duffel.
https://doi.org/10.1371/journal.pone.0346991.s002
(PDF)
S3 Fig. De rol van de beweeg- en eetomgeving in Herselt.
https://doi.org/10.1371/journal.pone.0346991.s003
(PDF)
S5 Fig. 1st presented CLDs during GMB session 2.
https://doi.org/10.1371/journal.pone.0346991.s005
(PDF)
Acknowledgments
We thank the municipalities of Duffel and Herselt for participating in this project. We thank all participants for their commitment, work and time.
References
- 1. Hruby A, Hu FB. The epidemiology of obesity: A big picture. Pharmacoeconomics. 2015;33(7):673–89. pmid:25471927
- 2.
WHO Regional Office for Europe. WHO European Regional Obesity Report 2022 [Internet]; 2022 [cited 2024 Jun 24]. Available from: https://www.who.int/europe/publications/i/item/9789289057738
- 3. Marques A, Peralta M, Naia A, Loureiro N, de Matos MG. Prevalence of adult overweight and obesity in 20 European countries, 2014. Eur J Public Health. 2018;28(2):295–300. pmid:29036436
- 4. Diderichsen F, Hallqvist J, Whitehead M. Differential vulnerability and susceptibility: How to make use of recent development in our understanding of mediation and interaction to tackle health inequalities. Int J Epidemiol. 2019;48(1):268–74. pmid:30085114
- 5. Carrillo-Álvarez E, Kawachi I, Riera-Romaní J. Neighbourhood social capital and obesity: A systematic review of the literature. Obes Rev. 2019;20(1):119–41. pmid:30306717
- 6. Christie CD, Consoli A, Ronksley PE, Vena JE, Friedenreich CM, McCormack GR. Associations between the built environment and physical activity among adults with low socio-economic status in Canada: A systematic review. Can J Public Health. 2021;112(1):152–65. pmid:32833139
- 7. Mackenbach JD, Nelissen KGM, Dijkstra SC, Poelman MP, Daams JG, Leijssen JB. A systematic review on socioeconomic differences in the association between the food environment and dietary behaviors. Nutrients. 2019;11(9):9.
- 8. Pitt E, Gallegos D, Comans T, Cameron C, Thornton L. Exploring the influence of local food environments on food behaviours: A systematic review of qualitative literature. Public Health Nutr. 2017;20(13):2393–405. pmid:28587691
- 9. Salvo G, Lashewicz BM, Doyle-Baker PK, McCormack GR. Neighbourhood built environment influences on physical activity among adults: A systematized review of qualitative evidence. Int J Environ Res Public Health. 2018;15(5):897. pmid:29724048
- 10. Sawyer ADM, van Lenthe F, Kamphuis CBM, Terragni L, Roos G, Poelman MP, et al. Dynamics of the complex food environment underlying dietary intake in low-income groups: A systems map of associations extracted from a systematic umbrella literature review. Int J Behav Nutr Phys Act. 2021;18(1):96. pmid:34256794
- 11. Townshend T, Lake A. Obesogenic environments: Current evidence of the built and food environments. Perspect Public Health. 2017;137(1):38–44. pmid:28449616
- 12. Dixon BN, Ugwoaba UA, Brockmann AN, Ross KM. Associations between the built environment and dietary intake, physical activity, and obesity: A scoping review of reviews. Obes Rev. 2021;22(4):e13171. pmid:33369097
- 13. Sallis JF, Cerin E, Kerr J, Adams MA, Sugiyama T, Christiansen LB, et al. Built environment, physical activity, and obesity: Findings from the international physical activity and environment network (IPEN) adult study. Annu Rev Public Health. 2020;41:119–39. pmid:32237990
- 14. Hoenink JC, Lakerveld J, Rutter H, Compernolle S, De Bourdeaudhuij I, Bárdos H, et al. The moderating role of social neighbourhood factors in the association between features of the physical neighbourhood environment and weight status. Obes Facts. 2019;12(1):14–24. pmid:30673683
- 15. Gebauer S, Schootman M, Xian H, Xaverius P. Neighborhood built and social environment and meeting physical activity recommendations among mid to older adults with joint pain. Prev Med Rep. 2020;18:101063. pmid:32140385
- 16. Zorbas C, Palermo C, Chung A, Iguacel I, Peeters A, Bennett R, et al. Factors perceived to influence healthy eating: A systematic review and meta-ethnographic synthesis of the literature. Nutr Rev. 2018;76(12):861–74. pmid:30202944
- 17. Darmon N, Drewnowski A. Contribution of food prices and diet cost to socioeconomic disparities in diet quality and health: A systematic review and analysis. Nutr Rev. 2015;73(10):643–60. pmid:26307238
- 18. Mazzucca S, Arredondo EM, Hoelscher DM, Haire-Joshu D, Tabak RG, Kumanyika SK, et al. Expanding implementation research to prevent chronic diseases in community settings. Annu Rev Public Health. 2021;42:135–58. pmid:33467924
- 19. Friel S, Pescud M, Malbon E, Lee A, Carter R, Greenfield J, et al. Using systems science to understand the determinants of inequities in healthy eating. PLoS One. 2017;12(11):e0188872. pmid:29190662
- 20. Karapici A, Cummins S. A participatory approach to model the neighbourhood food environment. PLoS One. 2024;19(1):e0292700. pmid:38266019
- 21. Carey G, Malbon E, Carey N, Joyce A, Crammond B, Carey A. Systems science and systems thinking for public health: A systematic review of the field. BMJ Open. 2015;5(12):e009002. pmid:26719314
- 22. Finegood DT. The importance of systems thinking to address obesity. Nestle Nutr Inst Workshop Ser. 2012;73:123–37; discussion 139-41. pmid:23128771
- 23.
Butland B, Jebb S, Parry V. Foresight. Tackling obesities: future choices. Project report. [Internet]; 2007 [cited 2020 Dec 4]. Available from: https://www.semanticscholar.org/paper/Foresight.-Tackling-obesities%3A-future-choices.-Butland-Jebb/e7fb5766dc0838ea5e163218214f65994ab4599f?p2df
- 24. Zheng L, Xue Y, Chen X, Ung COL, Hu H. Application of system dynamics approach in developing health interventions to strengthen health systems to combat obesity: A systematic literature review and critical analysis. BMC Public Health. 2025;25(1):1580. pmid:40301836
- 25. Hagenaars LL, Schmidt LA, Groeniger JO, Bekker MPM, ter Ellen F, de Leeuw E, et al. Why we struggle to make progress in obesity prevention and how we might overcome policy inertia: Lessons from the complexity and political sciences. Obesity Rev. 2024;25(5).
- 26. Morshed AB, Kasman M, Heuberger B, Hammond RA, Hovmand PS. A systematic review of system dynamics and agent-based obesity models: Evaluating obesity as part of the global syndemic. Obes Rev. 2019;20(Suppl 2):161–78. pmid:31317637
- 27. Siokou C, Morgan R, Shiell A. Group model building: A participatory approach to understanding and acting on systems. Public Health Res Pract. 2014;25(1):e2511404. pmid:25828443
- 28. Gerritsen S, Renker-Darby A, Harré S, Rees D, Raroa DA, Eickstaedt M, et al. Improving low fruit and vegetable intake in children: Findings from a system dynamics, community group model building study. PLoS One. 2019;14(8):e0221107. pmid:31415644
- 29. Gerritsen S, Harré S, Rees D, Renker-Darby A, Bartos AE, Waterlander WE, et al. Community group model building as a method for engaging participants and mobilising action in public health. Int J Environ Res Public Health. 2020;17(10):3457. pmid:32429183
- 30. Hendricks G, Savona N, Aguiar A, Alaba O, Booley S, Malczyk S, et al. Adolescents’ perspectives on the drivers of obesity using a group model building approach: A South African perspective. Int J Environ Res Public Health. 2022;19(4):2160. pmid:35206348
- 31. McGlashan J, Hayward J, Brown A, Owen B, Millar L, Johnstone M, et al. Comparing complex perspectives on obesity drivers: Action-driven communities and evidence-oriented experts. Obes Sci Pract. 2018;4(6):575–81. pmid:30574350
- 32. Savona N, Macauley T, Aguiar A, Banik A, Boberska M, Brock J. Identifying the views of adolescents in five European countries on the drivers of obesity using group model building. Eur J Public Health. 2021;31(2):391–6. pmid:33608719
- 33. Vennix JAM, Akkermans HA, Rouwette EAJA. Group model-building to facilitate organizational change: An exploratory study. Syst Dyn Rev. 1996;12(1):39–58.
- 34. Brennan LK, Sabounchi NS, Kemner AL, Hovmand P. Systems thinking in 49 communities related to healthy eating, active living, and childhood obesity. J Public Health Manag Pract. 2015;21:S55–69.
- 35. Calancie L, Fullerton K, Appel JM, Korn AR, Hennessy E, Hovmand P, et al. Implementing group model building with the Shape Up Under 5 community committee working to prevent early childhood obesity in Somerville, Massachusetts. J Public Health Manag Pract. 2022;28(1):E43–55. pmid:32810067
- 36. Nelson DA, Simenz CJ, OʼConnor SP, Greer YD, Bachrach AL, Shields T, et al. Using group model building to understand factors that influence childhood obesity in an urban environment. J Public Health Manag Pract. 2015;21 Suppl 3:S74-8. pmid:25828225
- 37. Guariguata L, Rouwette EA, Murphy MM, Saint Ville A, Dunn LL, Hickey GM, et al. Using group model building to describe the system driving unhealthy eating and identify intervention points: A participatory, stakeholder engagement approach in the Caribbean. Nutrients. 2020;12(2):384. pmid:32024025
- 38. Mui Y, Ballard E, Lopatin E, Thornton RLJ, Pollack Porter KM, Gittelsohn J. A community-based system dynamics approach suggests solutions for improving healthy food access in a low-income urban environment. PLoS One. 2019;14(5):e0216985. pmid:31086409
- 39. Guariguata L, Unwin N, Garcia L, Woodcock J, Samuels TA, Guell C. Systems science for developing policy to improve physical activity, the Caribbean. Bull World Health Organ. 2021;99(10):722–9. pmid:34621090
- 40.
Glickman AR, Clark JK, Freedman DA. 12 - Participatory modeling of the food system: The case of community-based systems dynamics. In: Peters C, Thilmany D, editors. Food systems modelling [Internet]. Academic Press; 2022 [cited 2024 May 22]. p. 257–83. Available from: https://www.sciencedirect.com/science/article/pii/B9780128221129000035
- 41. Baugh Littlejohns L, Hill C, Neudorf C. Diverse approaches to creating and using causal loop diagrams in public health research: Recommendations from a scoping review. Public Health Rev. 2021;42:1604352. pmid:35140995
- 42. Frerichs L, Lich KH, Dave G, Corbie-Smith G. Integrating systems science and community-based participatory research to achieve health equity. Am J Public Health. 2016;106(2):215–22. pmid:26691110
- 43.
Hovmand PS. Introduction to community-based system dynamics. In: Community based system dynamics [Internet]. New York, NY: Springer New York; 2014 [cited 2020 May 25]. p. 1–16. Available from: http://link.springer.com/10.1007/978-1-4614-8763-0_1
- 44. Whitfield GP, Carlson SA, Ussery EN, Watson KB, Berrigan D, Fulton JE. National-level environmental perceptions and walking among urban and rural residents: Informing surveillance of walkability. Prev Med. 2019;123:101–8. pmid:30878571
- 45. Pisman A. Het ruimterapport 2018: het ruimtelijk patroon van Vlaanderen ontrafeld = Spatial patterns in Flanders analysed in RURA 2018. ReM. 2020;11(4):4.
- 46.
ILVO. Duurzame ontwikkeling van peri-urbane gebieden. ILVO [Internet]; 2020 [cited 2023 Jun 8]. Available from: https://ilvo.vlaanderen.be/nl/nieuws/duurzame-ontwikkeling-van-peri-urbane-gebieden
- 47. Rijsbosch L. Ontwerpen met open ruimte in peri-urbane gebieden. Agora. 2016;32(3):38.
- 48. D’Hooghe S, Inaç Y, De Clercq E, Deforche B, Dury S, Vandevijvere S, et al. The CIVISANO protocol: A mixed-method study about the role of objective and perceived environmental factors on physical activity and eating behavior among socioeconomically disadvantaged adults. Arch Public Health. 2022;80(1):219. pmid:36199109
- 49. D’Hooghe S, Inaç Y, Deforche B, Van Dyck D, de Ridder K, Vandevijvere S. The role of the perceived environment for recreational walking among adults in socioeconomically disadvantaged situations: A study using walk-along interviews. SSM - Popul Health. 2023;23:101456.
- 50. Inaç Y, D’Hooghe S, Van Dyck D, De Clercq EM, Deforche B, Van de Weghe N, et al. Active travel as constrained necessity or choice: Individual and environmental factors among adults with a lower socioeconomic status in peri-urban Flanders, Belgium. Health Place. 2026;98:103602. pmid:41581324
- 51. D’Hooghe S, Inaç Y, Vandevijvere S, Deforche B, de Ridder K, Van Dyck D, et al. Exploring the multiple dimensions of perceived food access in the local food environment in Flanders: Perceptions of adults in socioeconomically disadvantaged situations. Appetite. 2024;203:107609. pmid:39094845
- 52. D’Hooghe S, Inaç Y, Vandevijvere S, Deforche B, de Ridder K, Van Dyck D, et al. Exploring the multiple dimensions of perceived food access in the local food environment in Flanders: Perceptions of adults in socioeconomically disadvantaged situations. Appetite. 2024;203:107609. pmid:39094845
- 53.
Glymour M, Avendano M, Kawachi. Socioeconomic status and health. In: Social epidemiology. Oxford, UK: Oxford University Press; 2014.
- 54. Montano D. Socioeconomic status, well-being and mortality: A comprehensive life course analysis of panel data, Germany, 1984-2016. Arch Public Health. 2021;79(1):40.
- 55. Laaksonen M, Sarlio-Lähteenkorva S, Lahelma E. Multiple dimensions of socioeconomic position and obesity among employees: The Helsinki Health Study. Obes Res. 2004;12(11):1851–8. pmid:15601982
- 56. Cundiff JM, Smith TW, Uchino BN, Berg CA. Subjective social status: Construct validity and associations with psychosocial vulnerability and self-rated health. Int J Behav Med. 2013;20(1):148–58.
- 57.
Rouwette EAJA, Vennix JAM. Group model building. In: Meyers RA, editor. Encyclopedia of complexity and systems science [Internet]. New York, NY: Springer; 2009 [cited 2024 May 1]. p. 4474–86. Available from: https://doi.org/10.1007/978-0-387-30440-3_264
- 58. Finegood DT, Merth TDN, Rutter H. Implications of the foresight obesity system map for solutions to childhood obesity. Obesity (Silver Spring). 2010;18 Suppl 1:S13-6. pmid:20107455
- 59. Richardson GP, Andersen DF. Teamwork in group model building. Syst Dyn Rev. 1995;11(2):113–37.
- 60. Barakat S, McLean SA, Bryant E, Le A, Marks P, National Eating Disorder Research Consortium, et al. Risk factors for eating disorders: Findings from a rapid review. J Eat Disord. 2023;11(1):8. pmid:36650572
- 61. Mukherjee K. Poverty as a cause and consequence of Ill health. Int J Epidemiol Res [Internet]. 2015 [cited 2024 Jun 28]. Available from: https://www.semanticscholar.org/paper/Poverty-as-a-cause-and-consequence-of-Ill-health-Mukherjee/a81ff52542fefaf92c8bf5a8172890d178a13806
- 62. Ter Ellen F, Oude Groeniger J, Stronks K, Hagenaars LL, Kamphuis CBM, Mackenbach JD, et al. Understanding the dynamics driving obesity in socioeconomically deprived urban neighbourhoods: An expert-based systems map. BMC Med. 2025;23(1):2. pmid:39762884
- 63. Stankov I, Useche A, Lekkas P, Henson RM, Headen I, Purtle J. Causal loop diagrams of food systems, diet, and obesity: A scoping review of methods, contextual factors, and outcomes. Obes Rev. 2025;26(11):e13966. pmid:40517776
- 64. Djojosoeparto SK, Sonneveld NA, Kamphuis CBM, Poelman MP. Why healthy eating is not equally attainable for all: A participatory systems perspective on dietary behaviour of people living on a low income in the Netherlands, and recommendations for systemic change. Arch Public Health. 2026;84(1):46. pmid:41606639
- 65. Camarero L, Oliva J. Thinking in rural gap: Mobility and social inequalities. Palgrave Commun. 2019;5(1):1.
- 66.
Marmot M. Fair society, healthy lives; 2013. p. 1–74.
- 67. Garrott K, Ogilvie D, Panter J, Petticrew M, Sowden A, Jones CP, et al. Development and application of the Demands for Population Health Interventions (Depth) framework for categorising the agentic demands of population health interventions. BMC Glob Public Health. 2024;2(1):13. pmid:39681925
- 68. Nieuwendyk LM, Belon AP, Vallianatos H, Raine KD, Schopflocher D, Spence JC, et al. How perceptions of community environment influence health behaviours: Using the Analysis Grid for Environments Linked to Obesity Framework as a mechanism for exploration. Health Promot Chronic Dis Prev Can. 2016;36(9):175–84. pmid:27670920
- 69. Swinburn BA, Sacks G, Hall KD, McPherson K, Finegood DT, Moodie ML, et al. The global obesity pandemic: Shaped by global drivers and local environments. Lancet. 2011;378(9793):804–14. pmid:21872749
- 70. Bauman AE, Reis RS, Sallis JF, Wells JC, Loos RJF, Martin BW, et al. Correlates of physical activity: Why are some people physically active and others not? Lancet. 2012;380(9838):258–71. pmid:22818938
- 71. Rutter H, Savona N, Glonti K, Bibby J, Cummins S, Finegood DT, et al. The need for a complex systems model of evidence for public health. Lancet. 2017;390(10112):2602–4. pmid:28622953
- 72. Beard E, Lorencatto F, Gardner B, Michie S, Owen L, Shahab L. Behavioral intervention components associated with cost-effectiveness: A comparison of six domains. Ann Behav Med. 2021;56(2):176–92.
- 73. Sterman JD. Learning in and about complex systems. Syst Dyn Rev. 1994;10(2–3):291–330.