Application of system dynamics approach in developing health interventions to strengthen health systems to combat obesity: a systematic literature review and critical analysis

Abstract

Background

Obesity is an escalating global public health challenge that is expected to impact a significant portion of the world's population in the coming decades. It leads to severe health conditions such as diabetes, cardiovascular diseases, and cancer, imposing significant economic burdens on health systems. Traditional intervention strategies, which emphasize individual lifestyle changes, fail to address the complex, systemic nature of obesity. This study aims to systematically review the application of system dynamics modelling (SDM) in obesity control, focusing on analyzing modelling methodologies and conducting a quality assessment of the included studies.

Methods

Employing a comprehensive systematic literature retrieval, we explored terms pertinent to overweight/obesity and system dynamics across three databases, including PubMed, Web of Science, and Scopus. This search culminated in identifying peer-reviewed studies published from the inception of these databases until July 2024. Quality assessment was used to evaluate the SDM for obesity control. The protocol of this systematic review has been registered on PROSPERO (CRD42024554520).

Results

Thirty studies were identified through a systematic review. These studies primarily focus on the effects of SD approaches, such as individual lifestyle changes, policy interventions within populations, and socio-economic and environmental improvements on obesity control. Among them, eleven studies completed the entire SDM process. Twenty-seven studies presented conceptual models, of which twenty-five developed casual loop diagrams (CLD). Seventeen studies conducted computational system dynamics modelling, with thirteen constructing stock-flow diagrams (SFD). Additionally, fourteen studies performed simulation analyses. These models facilitated multi-level strategies to reduce obesity prevalence.

Conclusions

Using SDM approaches has significant potential to enhance the effectiveness of obesity interventions and optimize resource allocation. Our study into the application of SDM in the design of obesity health interventions revealed its ability to promote multi-level, cross-sectoral cooperation and coordination, thereby enhancing the effectiveness of interventions. Further exploration and optimization of obesity health interventions can significantly advance health systems and welfare.

Background

Obesity is a pervasive and pressing public health issue that poses a significant challenge to health systems worldwide nowadays [1, 2]. It is projected that the global prevalence of obesity will surge from 14% in 2020 to 24% by 2035, affecting nearly two billion adults, adolescents, and children [3]. Obesity serves as a principal driver of numerous diseases, including metabolic disorders and a range of non-communicable diseases such as diabetes, cardiovascular diseases, and cancer, thus imposing a substantial burden on public health infrastructures [4, 5]. Additionally, obesity incurs significant direct medical costs and indirect costs [6,7,8]. The economic impact of overweight and obesity is expected to grow from 1.96 trillion USD in 2020 to 4 trillion USD by 2035 [4]. This phenomenon is evident in both developed and developing countries, highlighting the universality and severity of the obesity crisis [9].

Traditionally, the strategies for preventing obesity focused principally on changes in the lifestyles of individuals, particularly dietary habits and physical activity levels. This approach ignores the complex and systemic causes of obesity [10]. However, the emergence of the obesity crisis shows a strong relation with socio-economic position, dietary culture, and the food environment. All these factors together have resulted in an unprecedented rise in the cases of obesity both in developed and developing countries, which further complicates the task of responding to this global health problem [11, 12].

In 2019, the Lancet Obesity Commission emphasized the importance of systems science approaches in obesity research [13]. Such an endorsement underscores the need to employ methodologies that can unravel the complexities of obesity, making system dynamics particularly pertinent. System dynamics modeling (SDM), an integral component of systems science, dissects the intricate interdependencies among variables within a system and their resultant nonlinear behaviors [14]. Using simulation to anticipate the outcomes of interventions on complex health dilemmas, SDM provides a distinctive perspective for analyzing and tackling the issue of obesity [15]. These models effectively capture the complexity and multiple factors involved in obesity, showing how it develops over time [16]. In addition, they help simulate both the long-term and delayed effects, making it easier for policymakers to predict the outcomes of different obesity interventions more accurately [17, 18]. Addressing obesity necessitates comprehensive intervention strategies that span individual behavioral modifications, policy development at the societal level, and enhancements to the environmental backdrop [19,20,21]. SDM facilitates a deeper understanding of cross-level interactions, evaluates the cumulative impacts of diverse interventions, and thus underpins the development of robust, scientifically-backed strategies for comprehensive obesity management [22].

Nevertheless, the current literature lacks a thorough analysis and evaluation of the methodological approaches of SDM in obesity health interventions. Therefore, this study aims to systematically review the application of SDM in developing health interventions to strengthen health systems against obesity through analyzing modeling methodologies. Specifically, this study has three objectives. First, this study aims to explore common methodologies for constructing system dynamics models for obesity. Second, the study aims to analyze the key feedback loops and variables contributing to obesity and examine the simulation outcomes of obesity interventions. Third, the study aims to assess the quality of SDM applications in obesity and propose an evaluation framework for assessing the quality and appropriateness of individual SDM applications. The ultimate goal will be to advance the understanding and practice of SDM in health intervention development and assessment of obesity to improve health systems and prevent and manage obesity.

Methods

This review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines [23]. Throughout the review process, we systematically identified, summarized, and thoroughly analyzed scholarly papers that develop or apply SDM to prevent and control obesity. The protocol of this systematic review has been registered on PROSPERO (CRD42024554520).

Search strategy

We conducted searches on electronic databases including PubMed, Web of Science, and Scopus. We used the following keywords in the title, abstract, and keyword fields to ensure the capture of literature closely related to the target theme: (Obesity OR "Weight Loss" OR Overweight OR Adiposity OR obes* OR Bodyweight OR "Body Weight") AND "Systems Analysis" OR "Nonlinear Dynamics" OR "Computer Simulation" OR ((system dynamic OR system dynamics) AND (model* OR simulat*)) OR nonlinear dynamics OR dynamics simulation model OR causal loop OR causal loops OR (stock* AND flow* AND (model* OR simulat*)) OR "system dynamic" OR "system dynamics" OR "system* modeling" OR "system thinking" OR "system science" OR "system approach" OR "system theory"). Additionally, to further enhance the accuracy and completeness of our retrieval, we tailored our search strategies to each database's specific features. Beyond keyword-based retrieval, we employed a "snowballing" technique, identifying more relevant articles through citation and author searches. This method helped us uncover literature that may not have been captured through keyword searches.

Literature selection

The following review strategies were adopted for this study to be certain that the selection of research literature met established criteria. Initially, the titles and abstracts of all publications retrieved through the searches were independently reviewed by two researchers (LZ and YX). Subsequently, these two researchers thoroughly reviewed the full texts of the articles that passed the initial screening to confirm further whether each article met all eligibility criteria. Specifically, articles had to meet the following conditions to be included in the study. First, the article must be published in a peer-reviewed academic journal to ensure its scholarly quality and reliability. Second, the article must be written in English to facilitate in-depth reading and analysis by the research team. Third, the articles should employ system dynamics modeling techniques such as causal loop diagrams, stock and flow diagrams, or simulation, which are central methodological requirements for this study. Fourth, the article's theme must be closely related to obesity prevention or control, aligning with the core research objectives of the study. Fifth, the article's publication date should be between the establishment date of the databases and July 2024 to ensure the timeliness and completeness of the research. Articles meeting the above criteria were organized and saved into the EndNote reference management software program to facilitate subsequent data extraction and analysis.

Data extraction

We developed a standardized data extraction form. First, it included general characteristics such as author(s), year, country, setting, population of focus and study aims. Second, it addressed model development, including data sources, involvement of stakeholders, data analysis methods, tools for data analysis and diagramming. Third, it covered the model characteristics and structure, including the SDM methods, diagramming, subsystems, feedback loops, main model variables, stocks and flows. This structured form was designed to facilitate the thorough and systematic extraction of relevant data from the included studies, ensuring consistency and comprehensiveness in capturing critical information necessary for analysis and synthesis.

Quality assessment

To evaluate the quality of system dynamics modeling and documentation, this study developed nine quality criteria based on three key references [24,25,26], covering research objectives, assumptions, design, validation, data sources, and result interpretation to ensure that the models effectively capture the complexity and real-world impact of obesity interventions. Scores were assigned to each criterion (0–5), and the total score reflected the overall quality of the study. A detailed description of the assessment process is provided in the Supplementary File (Table S2).

These criteria offer a comprehensive evaluation framework for SDM in obesity interventions. Obesity is a complex public health issue influenced by multiple factors, including lifestyle, community engagement, and socioeconomic conditions [27]. System dynamics models are well-suited to simulate the long-term effects of interventions through dynamic features such as feedback mechanisms and delay effects [28]. Therefore, clarity in research objectives and explicit assumptions ensures that the model accurately reflects the key drivers of obesity and aligns with real-world interventions [29]. Clear model design and structure transparency help researchers understand and apply the model, while validation and data reliability ensure the scientific rigor and applicability of the model in addressing obesity issues [30]. Moreover, the presentation and interpretation of results provide strong support for policymakers in designing targeted and effective obesity strategies [31]. Thus, these criteria not only provide a scientific basis for evaluating system dynamics models in obesity interventions but also ensure their operability and relevance.

Results

General characteristics of included literature

Through electronic searches, 10,745 records were identified, from which 749 duplicates were removed, leaving 9,996 records for preliminary screening. In a two-step process for screening the identified records, each record’s title and abstract were first assessed to determine if they met the inclusion criteria. Records whose titles and abstracts did not provide sufficient information for determination were temporarily retained for further review. After the initial screening, the full texts of 208 records were evaluated. During this secondary screening, 178 articles that did not meet the inclusion criteria were excluded, ultimately resulting in 30 studies being included in the review (see Fig. 1).

Fig. 1

PRISM diagram of identification and selection of studies for the systematic review

All 30 selected studies were published between 2011 and 2024 (see Table 1). The majority of these investigations were conducted in developed countries, with 10 studies taking place in the United States [32,33,34,35,36,37,38,39,40,41]. The settings of the studies were diverse, encompassing both general and specific locales, including four studies [33, 39, 42, 43] that focused on particular communities, three [44,45,46] on schools, and one on a household setting [47]. Thirteen studies exclusively examined children [33, 39,40,41, 43, 44, 46, 48,49,50,51,52,53,54], five focused solely on adolescents [44, 45, 55,56,57], three targeted both children and adolescents [35, 36, 47], three investigated adults [34, 58, 59], and two focused on children and their families [38, 40]. Additionally, four studies addressed broader groups, such as community members [42], adults and youth [37], and the general population [60, 61]. The principal study aims of these studies were to apply system dynamics conceptual modelling to explore the systemic causes of obesity in specific populations within particular contexts and to analyze or predict the effects of obesity interventions through SD computational modelling and simulation.

Table 1 General characteristics of the included studies

Model development

Model development is crucial for understanding how different data sources, stakeholder involvement, and analysis methods influence the SDM models in obesity interventions. Table 2 presents the data sources included in the studies, the involvement of stakeholders at different stages, and the methods and tools used for data analysis and model construction.

Table 2 Model development of the included studies

In the realm of conceptual modelling, the majority of studies predominantly employ a hybrid approach utilizing both primary and secondary data sources. Primary data sources include direct stakeholder engagement [42, 43], Group Model Building (GMB) sessions [38,39,40, 43,44,45, 49, 55, 56], stakeholder workshops [47, 58], and interviews [51, 54, 55]. Secondary data sources are primarily literature reviews, with some studies specifically citing established datasets such as the HBSC [48], NSW [50] datasets, and Census data [35, 50, 52]. In contrast, data sources used for deriving parameter values in computational models tend to be quantitative, frequently relying on specific datasets and databases. These include historical trends of diseases, cross-sectional surveys, longitudinal survey, census data, and peer-reviewed publications [34, 36, 37, 52, 53, 58, 60]. Some studies use cohort study data and existing U.S. surveillance data to inform their models [41, 59]. This indicates the important role of quantitative data in accurately predicting the effects of obesity interventions and the impact of policies.

Twenty-three studies involved the participation of relevant stakeholders. In six of these studies [32, 39, 42, 44, 55, 58], stakeholders were engaged in constructing causal loop diagrams (CLD). Six studies [33, 38,39,40, 43, 45] saw stakeholder participation in GMB sessions. Six studies [35, 41, 46, 47, 50, 57] involved stakeholders in model building and development. Stakeholders participated in model validation in five studies [32, 36, 47, 48, 60], while two studies [37, 56] involved stakeholders in project evaluation. Additionally, one study [49] was associated with model design, another study [51] with data collection, and one study [32] focused on formulating questions. The involvement with diversity underlines the importance of stakeholders throughout the research cycle, from the designing through to evaluation, and further underlines the relevance and applicability of the results.

Data analysis methodologies employed in the studies include: statistical analysis; qualitative analysis (including but not limited to analyses based on the GMB session and interviews); sensitivity analysis; and mixed-methods research. During the research, data analysis, and charting tools to support these analytical techniques will be seriously applied throughout. These range from system dynamics modelling and simulation software, such as Vensim [46], mapping/chart creation software such as STICK-E and KUMU [55], health impact prediction software such as PRISM [37], to analytical software like iThink®, SAS, and R with a "deSolve" package [50, 58]. Moreover, special online applications and software were also used, such as Online Silico App [58], NutriMod [60]. Many studies adopt participatory research approaches, particularly emphasizing the use of GMB techniques and Vensim software for the analysis and development of CLD, alongside deep stakeholder engagement (see Table 2).

Model characteristics and structure

Model characteristics and structure demonstrate how SDM is applied to represent the dynamic processes in obesity formation and interventions. Table 3 summarizes the SDM methods used, diagramming techniques, subsystems, feedback loops, and the main variables driving the obesity systems.

Table 3 Model characteristics and structure of the included studies

The process of system dynamics modeling is typically divided into three stages: conceptual model, computational model, and simulation [62]. In the conceptual model stage, the primary focus is to identify and analyze the causal relationships between variables in the system by building causal loop diagrams (CLD). This stage focuses on qualitative description and theory building. During the computational model stage, the causal relationships in the conceptual model are turned into mathematical formulas, creating stock and flow diagrams (SFD). This step changes the qualitative model into a quantitative one, allowing for more accurate analysis and predictions. At the simulation stage, the computational model is used to simulate the system's evolution, test different scenarios, and predict the future state of the system [63].

Among the thirty studies included, eleven studies completed the full process of system dynamics modeling [34, 36, 41, 46, 48, 50,51,52,53,54, 59]. Of these studies, twenty-seven papers presented conceptual models, with twenty-five studies constructing causal loop diagrams (CLD). Additionally, seventeen studies conducted research on computational system dynamics models, and thirteen of them built stock and flow diagrams (SFD). Furthermore, fourteen studies conducted simulations to test the effectiveness of obesity interventions at different levels.

Subsystems encompass a wide array of factors, primarily including individual behaviors and familial influences, psychological and cultural educational elements, health policies and community engagement, as well as environmental and socioeconomic factors. At the level of individual behaviors and familial impacts, the focus primarily revolves around health and health-related behaviors, partnership dynamics [38,39,40], and interactions between adolescents and their parents [55]. Regarding psychological and cultural educational factors, they include mental health, unhealthy dietary habits, body image, motivational aspects of physical exercise, traditional societal norms [45], and school-based nutritional education [48]. This layer of subsystems provides a more profound understanding of the potential social and psychological factors contributing to obesity. Concurrently, in the realm of community involvement and policy intervention, the emphasis lies on community empowerment, advocacy for policy and systemic changes [32], and policies promoting healthy diets and active lifestyles [38,39,40]. In terms of environmental and socioeconomic factors, some studies focus on the food environment [32, 37, 55, 57], physical activity settings [35, 43, 50, 53,54,55, 57], online environments [55, 57], fiscal and human resources [49], healthcare, and urban contexts [56], all illustrating how dynamics factors influence individual health choices and the prevalence of obesity (see Table 3).

Main variables and feedback loops driving obesity

A complicated interaction of lifestyle, psychological, and socioeconomic factors drives the increasing prevalence of obesity. Several studies have explored these dynamics using system dynamics models. Eight feedback loops connecting adolescent obesity with lifestyle habits, psychological conditions, and dietary behaviors were identified by Romanenko et al. [48], emphasizing how insufficient physical activity and an unhealthy eating pattern contribute to obesity, which in turn aggravates these behaviors and elevates psychological stress. Hendricks et al. [44] went further to identify five important feedback loops within South African adolescents, with emphasis on physical activity, unhealthy food intake, social media, and socio-economic status, pointing toward the multifactorial nature of obesity. Savona et al. [45] used the GMB to generate three feedback loops. These loops connected food advertising, mental health, and the influence of social media on adolescent obesity. The study covered five European countries and suggested a preference for multi-level prevention strategies. Swierad et al. [33] studied childhood obesity in Manhattan’s Chinatown. They explored how sociocultural traditions, family impact, and time spent indoors are interrelated. Their findings highlight the need for culturally sensitive interventions.

Other studies further illustrate the community and education factors contributing to obesity. Nelson et al. [38] and Brennan et al. [39] focused on how community environments contribute to childhood obesity, emphasizing factors like lack of access to healthy food, limited physical activity opportunities, and socioeconomic challenges within communities. Keane et al. [40] identified 27 factors affecting obesity by focusing on the impact of community cultural pride and youth participation in healthy behaviors. In another study, Lan et al. [46] made use of system dynamics modeling to investigate how personal lifestyle, family influence, school education, and peer interaction affect children's BMI, presenting health concepts and education as having an important influence on the change of children's BMI. Moreover, Carrete et al. [51] examined how community and school education impact childhood obesity, highlighting the lack of sports facilities in communities and insufficient healthy food education in schools (see Table 3).

Simulated obesity interventions and outcomes

System Dynamics (SD) models are practical tools for simulating obesity interventions, highlighting the multidimensional and interdisciplinary approaches required to address obesity [64]. These interventions range from lifestyle changes to community and government initiatives aiming to reduce obesity rates and improve health systems.

Lifestyle and community-based interventions have proven to be effective in reducing obesity rates. Romanenko et al. [48] predicted more than 8% reduction in adolescent obesity across Europe by 2026 through targeted measures such as increasing physical activity, fruit intake, reducing life dissatisfaction, and alleviating academic stress. Calancie et al. [32] employed a community-driven strategy, enhancing nutrition programs, food security coordination, and community empowerment, significantly improving public health. Guariguata et al. [58] focused on the Caribbean, where combined upstream (dietary improvements) and downstream (targeting high-risk populations) interventions were projected to reduce obesity by 3.4% by 2050. Shahid and Bertazzon [52] explored the impact of neighborhood walkability on childhood obesity in Calgary, finding that increasing walkability scores could noticeably decrease obesity rates. Abidin et al. [53, 54] used SDM to simulate the effects of dietary behavior changes on childhood obesity in the UK, showing that while interventions could reduce obesity rates, a longer timeframe is needed to meet governmental targets.

Economic and public health interventions have also demonstrated significant potential in addressing obesity. Chen et al. [34] found that a 16.6% reduction in obesity could be done through improving income mobility and increasing employment opportunities. Roberts et al. [50] have predicted that a close to 5% reduction in obesity rates by 2025 would be reached by enhanced public health programs and improved environmental facilities, based on nine simulations of childhood obesity-targeted interventions. In another study, Crielaard et al. [59] on health awareness and social norms found that health awareness could bring modest weight reductions, but adding social norms greatly improved outcomes, especially for men. Liu et al. [36] examined the imposition of a tax on sugary drinks and the recycling of its revenue to fund public health programs, showing that such policies do reduce the rate of obesity.

Comprehensively, multifaceted interventions that involve cooperation between government, community, and families are crucial for effective obesity reduction. For instance, Powell et al. [35] were able to demonstrate that instituting a suite of policies at the same time could reduce the rate of obesity from 18 to 3%. In accordance with Carrete et al. [51], family and government cooperation will be able to substantially lower childhood obesity rates with a combined approach over 25 years. Kuo et al. [37] used the PRISM model with comprehensive strategies focusing on healthy eating, active lifestyles, and health marketing to predict dramatic reductions in obesity and low physical activity rates by 2040. Frerichs et al. [41] showed, with a model of social transmission between adults and children, that reducing social transmission would directly lead to reductions in childhood obesity rates.

These interventions not only target individual behavioral changes but also involve structural adjustments at the social and environmental levels. By simulating the implementation effects and long-term impacts of various strategies, SDM helps researchers optimize intervention effectiveness and predict their comprehensive impact on public health, supporting the development of more effective obesity prevention and intervention strategies [65].

SDM quality assessment report

According to the quality assessment results presented in Table 4, the overall quality scores of the included system dynamics modeling (SDM) studies range from 29 to 37 points, with an average score of 32.6 points (out of a maximum of 45). In the various quality criteria, most studies excelled in the clarity of model design and structure, with an average score of 4.2; in the detailed description of the modeling process, the average score was 4.1. However, the comprehensiveness of model validation received the lowest score of 1.1. Regarding the clarity of research objectives and purpose, explicitness and appropriateness of model assumptions, and clarity and reliability of data sources, most studies scored between 3.6 and 4.0.

Table 4 The quality assessment scores of included system dynamics modelling studies

Discussion

This study systematically reviewed 30 papers that utilized SDM for obesity, focusing on the factors influencing obesity and the intervention measures. First, these SDM studies make it possible to dive into multifactorial reasons and complex interactions behind obesity. Most of these studies reflected the feedback loops of individual behaviors, socio-economic factors, and community influences considerations put together to deepen the complexity of the obesity trait. For instance, unhealthy dietary habits and insufficient physical activity not only directly lead to weight gain but also exacerbate obesity through their impact on mental health and social behaviors [48]. Second, this review explores how SDM can be used dynamically to address the obesity issue. System dynamics models can simulate the long-term effects of different intervention strategies that could show the potential effectiveness of multi-level and multifactorial interventions [66]. For example, they can test how well dietary practice, levels of physical activity, or policy measures work across varying socio-economic contexts and, therefore, be used to guide more precise and effective public health interventions [36, 37, 48].

However, it must be noted that SDM is a context-specific approach that needs to consider contextual factors carefully. In this review, several studies indicate that socio-economic and cultural differences in different contexts significantly impact the effectiveness of obesity interventions. For example, Romanenko et al. [48] found that public health infrastructure and education levels in different European countries influenced the effectiveness of interventions, while the studies by Hendricks et al. [44] and Savona et al. [45] showed different driving factors behind adolescent obesity in South Africa and Europe. Liu et al. [36] studied the policy of taxing sugary drinks and recycling the revenue for public health programs, but the effectiveness of such policies may vary depending on regional economic conditions and the strength of policy support. In resource-limited areas, these policies may face implementation challenges, so it is necessary to adjust the policy parameters in the SDM model based on the specific economic and social environment of the region to achieve the best results. These contextual differences emphasize the need for adjustments in SDM when implemented in different regions to fit the local socio-cultural and economic characteristics.

Regarding modeling research methods, conceptual models, visually represent the complex interactions within the obesity system. These models are crucial for identifying key intervention points and facilitating stakeholder collaboration, particularly in diverse socioeconomic and cultural contexts. For instance, the study of Swierad et al. [33] illustrated how reinforcing feedback loops in social norms, marketing of foods, and family configurations reinforce obesity. Computational models enable researchers to simulate and predict the long-term effects of specific interventions, such as economic policies or public health strategies [22]. Studies by Chen et al. [34] and Liu et al. [36] found that interventions in economics, such as improving income mobility and taxing sugar-sweetened beverages, could result in significant reductions in the rate of obesity. These models support policymakers in making decisions regarding the potential policy impacts and optimization of resource allocation through scenario simulations. The qualitative and quantitative methods integrated into SDM provide the systematic framework to understand obesity, which can help develop more relevant and effective public health interventions [67].

Most are, however, heavily reliant on secondary data sources from literature reviews and national health datasets that do not always reflect an accurate picture, hence their limitations in real-time conditions at the local and current levels. It is also found that most of the proposed models have not passed through proper calibration or validation with real-world data, which raises some concerns regarding the reliability of model outcomes once applied in practical settings [30]. Therefore, it is necessary for researchers to follow SDM research processes rigorously to ensure the validity of SDM.

However, as identified from studies such as Hendricks et al. [44] and Pinzon et al. [55], stakeholder involvement typically remains partial to some phases, mainly during the initial construction of CLDs, and does not necessarily involve continuous engagement right through the modelling and validation stages. This might limit the models in capturing evolving system dynamics or reflecting multiple stages of stakeholder perception in intervention planning. While GMB sessions stimulate group collaboration, most research, such as Nelson et al. [38] and Brennan et al. [39], have utilized single-session GMB processes devoid of follow-up iterations to enhance and refine models. This single session reduces the chances of dynamic interaction and incremental model improvement. Further research will have to consider iterative GMB, involving stakeholders at various stages in the development and validation of the model. This is how adaptive and more precise models and more effective policy interventions would be possible.

At present, research on obesity intervention utilizing SDM predominantly emphasizes behavioral modifications and socio-economic determinants. Nevertheless, the integration of clinical services—including nutritional counseling, behavioral therapy, pharmacotherapy, and surgical interventions—remains insufficiently represented within these models. Integrating clinical interventions into SDM can provide a more comprehensive understanding of obesity management [68]. By combining real-world clinical services with SDM, researchers can simulate personalized treatment plans and predict their long-term outcomes[69, 70], thereby developing more targeted strategies for obesity prevention and treatment, especially in addressing comorbidities like diabetes and cardiovascular diseases[58, 71].

With SDM applications continuing to expand into public health, the development of uniform scoring systems will further enable model comparison capabilities and ensure the validity of research findings [72]. Future research should further explore how clinical services can be integrated with community interventions, considering their application across different socio-economic conditions and cultural backgrounds. For example, in low-income areas, more attention may need to be given to resource availability and community involvement, whereas in high-income areas, the focus might be on lifestyle changes and health education. Through such adjustments, SDM can more effectively address the needs of obesity interventions across different contexts. Moreover, there is a need for future studies to focus on the long-term impact of clinical interventions in preventing obesity relapse and cost-effectiveness as a result of combining clinical services with community-based obesity prevention interventions. In fact, other studies will also be able to establish how such improvement of accessibility and quality of clinical services in either an underserved or rural area impacts obesity prevention strategies [73]. Through these investigations, the practical application of future SDM will be more closely aligned with real-world health systems, providing more robust support for developing targeted obesity interventions[74, 75].

To our knowledge, this is the first systematic review focusing on applying SDM to explore obesity, particularly in revealing the complex drivers of obesity and analyzing the long-term effects of obesity interventions. This review utilized a systematic search methodology covering multiple databases and strictly adhered to explicit inclusion and exclusion criteria. Our analysis of existing models, especially in applying multi-level intervention frameworks, lays a theoretical foundation for future research in the public health domain using SDM. However, there are certain limitations to this review. Firstly, we limited our analysis to studies published in English, which may have resulted in the exclusion of important non-English research and thus affected the comprehensiveness of our findings. Secondly, the heterogeneity among the studies included presented challenges in assigning a consistent score to the models. This variability may have led to significant discrepancies in quality during the model evaluation process, ultimately restricting our overall assessment of model validity. Lastly, due to the diversity in the studied SDMs, we were unable to establish a systematic framework connecting the contextual factors with model components. Future research in this area is highly encouraged.

Conclusion

The SDM approach has shown notable effectiveness in the development and analysis of obesity health interventions, particularly in reinforcing health systems. Our comprehensive examination of SDM's application in obesity prevention and control strategies highlights its ability to foster integration at the health system level and promote inter-sectoral collaboration. Furthermore, the introduction of SDM allows policymakers to examine the diverse factors influencing obesity from the initial stages of strategy design. These factors encompass lifestyle choices, community engagement, and socio-economic conditions. Such thorough consideration is crucial for optimizing resource allocation while enhancing the flexibility and adaptability of interventions. Accordingly, future research should continue to utilize SDM to improve model accuracy and the practical effectiveness of interventions, ultimately supporting the optimization of health systems and reducing the prevalence of obesity and its associated health complications.