An action‐oriented framework for systems‐based solutions aimed at childhood obesity prevention in US Latinx and Latin American populations

Summary Childhood obesity in US Latinx and Latin American populations is a persistent, complex public health issue and, as such, requires solutions grounded on systems science theory and methods. In this paper, we introduce an action‐oriented framework to design, implement, evaluate, and sustain whole‐of‐community systems changes for childhood obesity prevention in US Latinx and Latin American populations. Our framework covers six action steps: (1) foster multisectoral team; (2) map the system, its context, and drivers; (3) envision system‐wide changes; (4) effect system‐wide changes; (5) monitor, learn, and adapt; and (6) scale and sustain. We also propose 10 principles that put human and environmental rights and systems thinking at the center of these systems‐based solutions. For each action step, we provide a list of concrete activities, methods, approaches, and examples that can be used to guide and inform the work needed to achieve the expected outputs. Finally, we discuss how a wider adoption of systems science for childhood obesity prevention among US Latinx and Latin American populations can be encouraged and sustained.


| INTRODUCTION
It is well recognized that rising levels of childhood obesity in US Latinx and Latin American populations is a persistent, complex problem with important implications for health and beyond. It has the characteristics of a complex problem in that it is driven by influential factors and social actors (e.g., people and organizations) operating and dynamically interacting across multiple levels over time. [1][2][3][4] It therefore follows that childhood obesity prevention efforts will need to operate within this complex adaptive system and, as such, could attain better results with the application of systems science theory and methods. 5 Systems science is an interdisciplinary field engaged in the study of the properties of systems-integrated wholes made up by interdependent elements. Complex adaptive systems are special cases of systems: they are formed by many elements or components that are irreducibly entwined, interacting over time with no or low central coordination or control, and they create patterns and ways of functioning that are not displayed by the individual components and that adapt in response to changes in the context in which the system exists. 6 The levels of childhood obesity in US Latinx and Latin American populations can be conceptualized as the result of a complex adaptive system encompassing the food, school, and transport systems, sociocultural and environmental influences, and numerous other factors that interact to shape energy intake and expenditure. 4 For instance, Pérez-Escamilla et al. 1  Implementing changes to how a complex adaptive system is structured and operates can be understood as effecting a disruptiona threat, to some extent-to the current form and function of the system. Complex systems tend to settle in stable states and resist or overcome disruptions that push them toward a new regime. 7 That means that efforts to reconfigure the system tend to trigger the system's responses to the intervention itself, oftentimes resulting in failure to address the problem, an issue known as resistance to change (or intervention or policy resistance). 8 Calls to expand the application of systems thinking in efforts to prevent childhood obesity in US Latinx and Latin American populations have been increasing rapidly. However, much of the current literature is based on theory rather than practical applications. 2,9,10 There is limited guidance for interdisciplinary and multisectoral teams (e.g., public health professionals, professionals from other sectors, evaluators, academics, community leaders, decision-makers, and stakeholders) willing to develop system-based solutions to address childhood obesity in their communities.
In this paper, we introduce an action-oriented framework that addresses the need for a more concrete roadmap to incorporate and use systems science theory and methods to design, implement, evaluate, and sustain whole-of-community systems changes for childhood obesity prevention in US Latinx and Latin American populations.
In Table 1 you will find a glossary of systems science concepts present in the next sections. These terms are indicated in italics throughout the text.

| AN ACTION-ORIENTED FRAMEWORK FOR SYSTEMS CHANGES FOR CHILDHOOD OBESITY PREVENTION
Our action-oriented framework covers the activities that are critical to achieve sustained changes in systems that fuel and perpetuate childhood obesity in US Latinx and Latin American populations. In our framework (Figures 1 to 3 and Table S1), each activity is represented by a different action step (represented in the figures as "cogs") which together comprise the overall framework, as follows: creating and nurturing a multisectoral team (orange cog); understanding the system and the underlying causes of the problem (green cog); envisioning system-wide changes (blue cog); transforming vision into action through coordinated actions to change how the system operates (yellow cog); monitoring how the actions are implemented, understanding how the system as a whole and its various elements respond and adapting actions accordingly (purple cog); and identifying ways to scale and sustain system-wide changes (gray cog).
As illustrated in Figure 1, a system-based solution to reduce childhood obesity in US Latinx and Latin American populations is composed of the above activities that interact with each other over time T A B L E 1 Glossary of systems science concepts Adaptive capacity: capacity to adjust its own characteristics or behaviors to respond to existing and future conditions and problems. 11 Agent-based model: quantitative modeling technique that simulates a composition of autonomous entities (e.g., persons or organizations), called agents, that "make decisions" according to a set of behavioral rules that guide their interaction with the surrounding environment and other agents over time, shaping the system's global patterns. 12 Complex adaptive system: system formed by many components that are irreducibly entwined, interacting over time with no or minimal central coordination or control, creating collective patterns and ways of functioning that are not displayed by the individual components and that adapt in response to changes in the context the system exists. 6 Feedback loop: cause-and-effect chain that connects two or more factors in a circuit (or loop). Feedback loops can be positive (factors reinforce each other over time) or negative (factors balance each other over time). 13 Group model building: participatory method for involving people in a modeling process. 14 Just-in-time adaptive intervention: intervention design that adapts its properties (e.g., the type, timing, intensity) as conditions and contexts change, delivering support at the moment and in the context that it is most needed or is most likely to be receptive. 15 Network analysis: set of methods and techniques to describe and analyze networks, i.e., structures of relations or connections among entities (e.g., people, organizations, projects). 16 Self-organization: process through which a system's global structure arise solely from local interactions among the elements of the system, with no or minimal central coordination or control. 6 System-based solution: processes and actions to deal with a problem that are underpinned by systems thinking and systems science principles and methods. System dynamics model: quantitative modeling technique that uses coupled differential or integral equations to describe and analyze the global behavior of complex systems over time. These equations represent the factors and quantities involved in the systems and how they affect each other over time. 13 Systems science: interdisciplinary field engaged in the study of the properties of systems. 6 Systems thinking: way of thinking, conceptualize, and make sense of the world characterized by the application of core systems concepts (e.g., inter-relationships, feedback loops, adaptation, selforganization). 17 in direct and indirect ways and play a fundamental role in the success of the solution. Obstructions to the execution of any one activity, such as delaying action or depriving required resources, can cause negative ripple effects for the other activities and, consequently, impact the extent to which system-wide changes can be achieved and sustained.
Helping interdisciplinary teams develop system-based solutions to reduce childhood obesity in their communities is at the heart of our framework. To help with this goal, for each of the activities, we provide a non-exhaustive list of concrete actions ( Figure 2 and Table S1) and methods and approaches ( Figure 3 and Table S1) that can be used to guide and inform the work needed to achieve the expected In the next sections, we provide an overview of each action step of this framework.

| Foster a multisectoral team
A critical starting point for initiating system level change to prevent childhood obesity is to identify and meaningfully engage with the key people, organizations, and social actors who will play a role in understanding, changing, and sustaining local systems. For childhood obesity, this may involve engaging with a core team, and broader coalition, that includes leaders from the Latinx community, as well as parents, teachers, coaches, child health care providers, business, government, non-profit, and scientific leaders who have an interest in, or influence on, child wellbeing. 23 Without the careful identification of this breadth of stakeholders, the team is likely to have an incomplete perspective of the system and factors that impact child obesity, because people with differing perspectives and roles in system were not involved in articulating the problem and the underlying social determinants. 24 Engaging a diverse team is also needed to create a vision for the future; identify the best actions that are aligned with local priorities, capacity, and resources; and foster long-term buy-in and support needed for sustained community adoption.
Building this multisectoral team should begin by "starting where the people are" 25 and identifying key actors and trusted relationships that are committed to and/or engaged in the issue of childhood obesity. 23 Additionally, it is helpful to engage with people and organizations that play a gatekeeping or "bridging" role between health care or scientific team members and the community, 24 such as community organizers or liaisons who understand the Latinx community being served and are viewed as credible and trusted. 26 In addition, employing systematic methods for including the voices of community residents themselves, critical to enrich the information being gathered and acted upon, is been increasingly recognized. 27,28 The scientific members of the team should also be multidisciplinary, ideally integrating expertise in health and medicine, child development, social and cultural sciences, health policy, community psychology, data/analytic sciences, and systems science. Finally, a snowball engagement approach, 29 where initial team members and community leaders invite other stakeholders they know to get involved, can help to ensure there has been adequate reach. A useful step in fostering a multisectoral team is to map the network of relationships, such as existing collaborations, that exist among the members. Social network analysis is a useful tool for this 16 because it can provide insights into clusters of stakeholders that are already well connected or disconnected and identify important opinion leaders who are known by many in the team and who could provide leadership. 31 It also can be used to help make the coalitions more efficient, effective, and sustainable by employing network intervention strategies. 32 Intentional activities are needed to solidify the team as a collective who will work together on shared issues and goals. 24  A contentious point in the building of multisectoral teams for childhood obesity prevention initiatives is the participation of opposing forces, the food and beverage industry in particular, which has been involved in practices that increase children's risk for obesity and market unhealthy foods to communities with disadvantage, including Latinx, in favor of maximizing profit. 36 In our view, all actors who play a role in the form and function of the system should be identified and invited to productively contribute to the multisectoral team. However, those who fail to align with the team's vision, goals, culture, and environment may need to be disengaged from the team before the rela- Altogether, these activities can be used to foster a representative and multisectoral team of stakeholders who are engaged in creating systems change to address childhood obesity. These activities are intended to help foster a team and broader coalition that has a culture and environment that will encourage bi-directional instead of topdown relationships, and a space for collective knowledge creation, learning, and actions that will be essential qualities to support the team in their subsequent phases of action.

| Map the system, its context, and drivers
Central to a systems-based solution to reduce childhood obesity is to understand how factors at multiple levels (individual, interpersonal, organizational, social, environmental, economic, and policy-based) and actors across sectors operate and dynamically interact, shaping and Next is the mapping of factors, actors, processes, and contextual elements driving the levels of childhood obesity. 2,14,37 The goal is to draw a diagram of the system that explains the levels of childhood obesity as an endogenous consequence of the feedback structure between factors. 13,14,37 Causal loop diagrams are one of the most widely used tools for this end. 13  Task Force, and media), illustrating the interplay with and between subsystems and how they come together as a whole system. 9 It is important to acknowledge that knowledge about the system's form and function and the potential drivers of the levels of childhood obesity is distributed among the actors contained in the system. This

| Envision system-wide changes
Once the multisectoral team has a shared articulation of the problem and understanding of the system, its context, and putative drivers, the next step is to co-create a shared vision of the future. This action step is integral to the success of the entire initiative, given that without the true co-creation of a shared way forward, any suggested initiatives and actions taken are more likely to fail, suffer delays, and be less effective and unsustainable. This action step will benefit from drawing on the approaches and lessons in implementation science exemplified in this series. 41 Steps to co-creating a shared vision include (a) identifying potential these approaches in Section 2.6). [48][49][50] Public participation approaches (e.g., citizen panels) can be used to ensure that articulated actions are acceptable, feasible, effective, adaptable, and sustainable from the viewpoint of the population being served by the initiative.

| Effect system-wide changes
Interventions, such as those required to tackle childhood obesity in US Latinx and Latin American populations, can be envisioned as disruptions in complex adaptive systems. At this stage, the multisectoral team implements the system-wide changes that they have envisioned in the previous steps, and specific actions they have agreed upon, to achieve this change. These activities should be executed by the team and their collaborators, bringing in new expertise as needed, and coordinated through regular team meetings and ongoing stakeholder feedback. Team members' active engagement with broader community members and stakeholders throughout this phase will help to increase buy-in by the community being served, as well as the reach and sustainability of these actions.
Actions should be supported by evidence-based principles of dissemination and implementation science to accelerate and maximize their impact on the system and population being served by the intervention. Dissemination and implementation science identifies several strategies and factors that have been shown to increase the speed at which public health actions and interventions are adopted, and the extent to which they are effective and sustained. 51 Rogers's Diffusion of Innovations 52 is one widely used theory that provides useful insights into how to increase the adoption and spread of new ideas, practices, and behaviors in the population being served.
In public health interventions broadly, and childhood obesity interventions specifically, the adoption of new practices and health behaviors (e.g., healthier eating and physical activity habits) is often slow and not sustained. Diffusion of Innovations theory points to several characteristics of the innovation, of the "adopters", and of the system in which adoption is occurring, that can be addressed to help accelerate system-wide adoption. For example, system-wide changes to address childhood obesity in US Latinx and Latin American populations will more likely occur and be sustained if the programmatic actions (e.g., changes to school lunches) and individual behaviors (e.g., walking to school) to be adopted (i) are not too complex; (ii) are compatible with the communities' values and goals; (iii) can, when taken up by "early adopters", be observed by others (who may be contemplating adoption); and (iv) are something that people and organizations feel that they can "try out" (e.g., make an initial change for a short period of time). Focusing on getting opinion leaders among the stakeholders and in the community being served to be the initial adopters of these actions and behaviors can also increase adoption and buy-in by others.
The Reach, Effectiveness, Adoption, Implementation, and Maintenance (RE-AIM) framework also provides guidance to improve the translation of interventions and actions into effective system-wide change. 41 This includes strategies to increase the number and breadth of people in the population being served who are exposed to the actions (Reach), optimizing the positive health changes that happen ing and cumulative nature of evidence and knowledge generation required in systems science and, therefore, to be prepared to adjust the evaluation tools and processes as the system, and evaluative needs, changes over time.

| Scale and sustain
Intervention or policy resistance-the system's counter-response to an intervention, resulting in failure to address the target problem 8can be attributed mainly to the lack of understanding of the feedback loops triggered by our decisions and actions. 8 A central tenet to overcoming intervention or policy resistance and facilitating and sustaining the reconfiguration of the system is to preserve the resilience of desirable parts of the system while trying to overcome the persistence of harmful parts. 7 Part of this work needs to be the development and implementation of new policies that shift conditions that were holding a system in a state that resulted in negative outcomes, which can help to sustain change over the long term. But additional actions may need to be taken to facilitate the system's transition to, and the permanence of, the envisioned new form and function, ensure the incorporation of new habits and norms, and avoid unintended consequences and tendencies to return to the previous state.
One of the steps we suggest is to define adaptation tipping points and adaptation pathways, central elements of the Dynamic Adaptive Policy Pathways approach. 49,57 An adaptation tipping point specifies the conditions under which an action or a constellation of actions will no longer meet the objectives, and, therefore, new actions will be needed. 49,57,58 Adaptation pathways are a sequence of possible actions after a tipping point is reached. 49,57,59 These pathways can be developed and explored via scenario modeling (e.g., system dynamics and agent-based models) and/or more qualitatively using public participation approaches. The process of defining adaptation tipping points and adaptation pathways prompts those implementing system-based solutions to reduce childhood obesity in US Latinx and Latin American populations to consider and plan built-in triggers and mechanisms for adjustments to steer the course of action in response to intervention or policy resistance and sustain the system reconfiguration process.

| WAY FORWARD
Although awareness of and support for systems science in childhood obesity prevention among US Latinx and Latin American populations is growing, adoption of systems approaches among intervention researchers in childhood obesity has been slow. 62