The workshop’s main session comprised three panels of speakers who discussed the usefulness of systems science approaches for stakeholders in different sectors involved in the development of solutions to address obesity. Jamy Ard, professor in the Department of Epidemiology and Prevention and the Department of Medicine at the Wake Forest University Baptist Medical Center, moderated the first panel, during which speakers discussed engaging the business and private sector in systems science approaches. Sara Czaja, professor of gerontology in medicine at Weill Cornell Medicine, moderated the second panel, focused on communities. Stella Yi, assistant professor in the Department of Population Health at the New York University Grossman School of Medicine, moderated the third panel, which highlighted policy makers.
THE BUSINESS AND PRIVATE SECTOR: PROJECT PLAY
In the first panel, three speakers discussed a business case study of Project Play, an effort designed to address the U.S. obesity epidemic by building healthy communities through sport.
Tom Farrey, executive director of The Aspen Institute’s Sports & Society Program, reported that Project Play began with a conversation about how the space of sport activity (a component of the realm of physical activity) could be organized to help more youth be active on a regular basis so as to
make a meaningful contribution to addressing obesity in the United States. Project Play’s mission is to help stakeholders build healthy communities through sport, Farrey clarified, noting that this mission is sometimes misconstrued as an effort to simply grow sport participation. Although growing participation in sports is something that Project Play’s leaders desire, he explained, its agenda is not always fully aligned with the youth sport space. As an example, he pointed to the resumption of youth travel for sport tournaments relatively early in the COVID-19 pandemic, noting that Project Play’s perspective at that point was to postpone this kind of activity.
Project Play’s endeavor to organize the U.S. sport system to produce more optimal health outcomes began with a focus on children aged 12 and younger, Farrey recounted, with the aim of ensuring that every child has the opportunity to engage in structured or unstructured sport regardless of zip code or ability. During the remainder of his presentation, he described three steps that the initiative has taken in pursuit of that aim.
The first step, Farrey said, was to compile and organize existing research on the individual- and community-level benefits of physical activity and participation in sports. He shared graphics showing examples of lifelong health benefits for physically active youth (see Figure 3-1) and benefits for residents of active communities (see Figure 3-2). These graphics, he explained, are used in appeals to stakeholders to make physical activity and participation in sports a public priority.
According to Farrey, this research informed the development of a playbook called Sport for All, Play for Life, the nation’s first cross-sector “framework for action” for youth (aged 12 and younger) sport (The Aspen Institute, 2015). The playbook, he elaborated, was an outgrowth of 2 years of information gathering and roundtable discussions with hundreds of stakeholders and thought leaders, during which the best ideas were elicited and distilled into an attractive, easy-to-read document. The values of health and inclusion are integrated throughout the playbook, he observed, to highlight the goal of every child having an opportunity to be active through sport. He added that the playbook highlights eight sectors that touch children’s lives—national sport organizations, business/industry, technology/media, public health, community recreation, education, policy makers and civic leaders, and parents—and presents eight strategies (and 40 activation ideas) to apply in those sectors to encourage more children to be active through sport.
Farrey then reviewed the eight strategies, beginning with putting children at the center of designing sport experiences (see Box 3-1). Youth sports are usually designed by adults, he pointed out, but children’s voices can be incorporated by asking them what kinds of experiences they want. He noted that Project Play is developing electronic surveys to collect feedback from youth (a tactic used by video game companies) to inform the development of
SOURCES: Presented by Tom Farrey, September 16, 2020; The Aspen Institute, 2015. Figure used with permission of The Aspen Institute’s Project Play.
SOURCES: Presented by Tom Farrey, September 16, 2020; The Aspen Institute, 2020b. Figure used with permission of The Aspen Institute’s Project Play.
youth sport programs. The second strategy, Farrey continued, is to reintroduce free play. The third is to encourage sport sampling, which he described as promoting multisport play and introducing youth in underserved communities to a wide variety of sports, as well as pushing back on the trend of asking children to specialize in a single sport at an early age. The fourth strategy is to revitalize in-town, local sport leagues. According to Farrey, these sport structures are more accessible than travel teams, which he said tend to limit participation to high-income, two-parent families, yet they still deliver many of the benefits of sport. The fifth strategy is to think small, which Farrey illustrated with the example of brokering shared-use agreements to use community spaces creatively. The sixth strategy is to design for development, which means anchoring a sport system in the principles of developmentally appropriate play. This strategy, Farrey explained, ties into the seventh strategy, which is to train all coaches. He observed that U.S. youth sports have traditionally been administered by well-meaning but untrained volunteers who would benefit from learning how to make sport a positive experience for youth, such as by recognizing the differences in their physical, mental, and emotional capacities at different ages. The final strategy, Farrey said, is to emphasize prevention. He explained that this strategy emerged to address parents’ concern about their children sustaining physical and emotional injuries from participating in sport. Recognizing that barrier to participation led Project Play to realize that it should give parents the confidence that sport can be a health-building experience.
Farrey next shared an aspirational model for sustained participation in sport, where the priority for children up to at least age 12 is physical
literacy, emphasizing fundamental movement skills and a focus on the ability, confidence, and desire to be active. This foundation sets the stage for future sport participation, he said, regardless of whether that participation is at a recreational, competitive, or elite level.
Farrey went on to describe Project Play’s second step in pursuing its mission, which was mobilizing organizations. This effort began with the convening of a summit, at which then U.S. Surgeon General Vivek Murthy issued a call to action on the value of youth sport for promoting physical activity and public health. Since then, Farrey recalled, Project Play has conducted “state of play” audits that help communities assess where they are and how they could design investments and policies that would mobilize stakeholders to activate and improve the quality of play in their communities. Project Play also convenes stakeholders to share knowledge, Farrey continued, highlighting an annual summit that invites leaders from the eight sectors to share new tools, projects, and resources. In addition, the summit showcases the release of an annual State of Play report, which features trend data on sport participation in youth aged 6–12 and 13–17, stratified by sport, income, and gender.
Farrey next described the Project Play Summit 2020, which convened some of the largest and most influential organizations in the Project Play network to discuss how they could collectively invest in increasing youth activity levels. Although many of the organizations are competitors, Farrey remarked, they banded together to develop initiatives such as howtocoachkids.org, the nation’s first free resource aggregating coaching materials for any organization to use with its coaches. Another initiative was the Healthy Sport Index,1 he continued, which helps parents and stakeholders assess the relative risks and benefits of participating in the 10 most popular sports for high school students. He explained that an expert committee developed the index by aggregating data on each sport’s level of physical activity, safety (based on injury rates), and psychosocial development, and then evaluated and weighted the data to create an interactive dial that stakeholders can use to assign the relative importance they place on each of the three factors and receive suggestions for which sports might be most suitable for individual adolescents.
Farrey then described the final step—mobilizing parents. He highlighted a media campaign introduced in August 2019 called “Don’t Retire, Kid,” launched with the help of Kobe Bryant and ESPN, which directed parents to advice and resources related to maintaining youth participation in sports. Research suggests that most children play sports for an average of 2.9 years, he noted, and typically stop by age 11 (The Aspen Institute, 2019).
Bruce Y. Lee, professor of health policy and management at the City University of New York Graduate School of Public Health and Health
___________________
1 See https://healthysportindex.com (accessed November 4, 2020).
Policy and executive director of Public Health Informatics, Computational, and Operations Research (PHICOR), began his portion of the presentation by reiterating that the initial premise of Project Play is straightforward in that it aims to tackle the epidemic of youth physical inactivity and also increase youth participation in sports by working with communities. However, because the scope of the physical inactivity problem is complex, he asserted, it warrants a highly energetic, organized, and coordinated operation like Project Play. According to Lee, it is important for decision makers to understand the magnitude and scope of a problem so that they can rank it relative to other priorities. He pointed out that decision makers benefit from information that quantifies the impact of a problem, as well as the potential value of different types of interventions, but that providing this information is a challenge when the problem is complex.
Lee described a suite of systems science models called the Virtual Population Obesity Prevention (VPOP) Labs (developed by PHICOR), highlighting their utility for helping decision makers grapple with the complexity of problems related to diet or physical activity and their potential solutions. In these agent-based models, which he likened to a SimCityTM for obesity prevention, each person is represented by a computational agent whose actions may be influenced by others but are ultimately determined by individual freedom of choice (a concept known as autonomous decision making). Agents also exhibit basic complex adaptive behavior and learning, he added, which means they can learn from the past or do things differently depending on previous events.
VPOP uses a synthetic population that was built using U.S. Census data, Lee continued, and each agent has characteristics that include age, gender, race/ethnicity, socioeconomic status, home assignment, school assignment, height, and weight. The initial model was built for Baltimore, he noted, and the same structure then informed model building for additional cities, as well as the entire United States. He explained that modelers can assign the agents a daily schedule of activities, behaviors, and decisions that mimic a real person’s patterns, and each agent is also assigned a personalized metabolic model. The metabolic model attempts to represent what happens within the human body when calories are ingested, he elaborated, which can result in a daily weight increase or decrease (depending on caloric expenditure), as well as a change in body mass index (BMI) over time.
Lee reported that his team attached the metabolic model and agent characteristics to an embedded clinical and economic outcomes model representing key health conditions potentially associated with changes in BMI (Fallah-Fini et al., 2017). In this model of clinical and economic outcomes for different BMIs over a lifetime, each agent had probabilities of developing diabetes, cardiovascular disease, or cancer over time, he elaborated, and accrued varying levels of medical costs and productivity losses based
on the health condition(s) that developed and the health care resources that the agent consumed. He added that this work led to a general model representing all youth in the United States, which simulated the economic and health impact of increasing physical activity in that segment of the U.S. population (Lee et al., 2017).
Lee reported that at present, fewer than one-third of U.S. youth maintain the Sports & Fitness Industry Association’s recommended “active to a healthy level” (i.e., 25 minutes of physical activity three times per week). Results from Lee’s model suggested that increasing this number to at least 50 percent of the current youth population would avert billions of dollars in medical care costs and lost productivity. This savings could then eventually occur for every group of children that reached the 8–11 age range (see Figure 3-3). The model was also used to explore the effect of varying the type of physical activity, Lee added, and showed that the more vigorous the physical activity, the greater was the amount of savings. To illustrate this point, he stated that if 75 percent of U.S. children aged 8–11 were to reach the goal of “active to a healthy level,” $16.6 billion in direct medical costs and $23.6 billion in lost productivity could be averted. If all U.S. children in this age group achieved that goal, overweight and obesity in youth would decrease by 15 percent, and $26.3 billion in direct medical costs and $36 billion in lost productivity would be averted.
The model’s results are shared with stakeholders, Lee said, to make the point that getting more youth active can avert billions of dollars in medical costs and lost productivity and save millions of years of life. He asserted that even achieving the Centers for Disease Control and Prevention’s (CDC’s) recommended physical activity levels among 100 percent of youth is “not unachievable” because the activity level threshold is not high, and achieving this goal could result in averting $51.5 billion in direct medical costs, $69 billion in productivity losses, and saving 37.5 million years of life (Lee et al., 2017). These savings, he noted, including fewer youth with overweight and obesity, were calculated for specific geographic regions of Project Play so that the potential benefits could be tailored to those regions with significant disparities in access to coaches, facilities, mentors, and financial resources. He highlighted particular reports from southeast Michigan, western New York, and the greater Rochester and Finger Lakes region (Ralph C. Wilson, Jr. Foundation, 2020).
The panel’s third speaker, Dev Pathik, chief executive officer and founder of the Sports Facilities Companies, described his company’s mission to improve the health and the economic vitality of communities. The company pursues this mission, he elaborated, by serving as a trusted resource for communities that want to plan, fund, develop, or manage youth and amateur community-based sport and recreation facilities. He explained that the Sports Facilities Companies advises municipalities and
NOTES: CDC = Centers for Disease Control and Prevention. For people aged 6 years and above, being physically “active to a healthy level,” as defined by the Sports & Fitness Industry Association, is defined as 25 minutes of high-calorie-burning physical activity three times per week. For those aged 6–17, CDC recommends at least 60 minutes of moderate physical activity each day.
SOURCES: Presented by Bruce Y. Lee, September 16, 2020; Lee et al., 2017. Figure used with permission from Project Hope.
large institutions on infrastructure planning and resource allocation so that they can activate their objectives around sport and recreation using a market research–driven process. According to Pathik, as the organization has recognized the significance of the impact of sport with regard to positive health and social outcomes across the lifespan, greater emphasis has been placed on considering lower-income populations and providing universal access to sport and recreation. Pathik added that the work of his organization is a good example of how practitioners can react to a systems science approach. He noted that the organization has revised its approaches as a result of its engagement in the work of Project Play and the systems science modeling described earlier in the presentation by Lee.
Pathik next reviewed five access factors for healthy communities, emphasizing that infrastructure plays a role in these factors, which are intended to improve health and social outcomes for youth. The first factor is safe and free spaces to play, such as a safe neighborhood where youth can move about and engage in recreational activities. A second factor is activity-focused urban design, which Pathik explained using example measures such as the interconnectivity of recreation and sport assets and walkability. A third factor he cited is the utilization and programming of school assets and sport and recreation facilities. A fourth is investing in multigenerational programs and events. Finally, the fifth factor is multimodal transportation. According to Pathik, these five access factors help communities think more broadly about how they participate in, contribute to, and fund or use partnerships to promote wellness among their residents.
Pathik noted that the last recession brought significant defunding of parks and recreation, which prompted the private sector to capitalize on opportunities in that marketplace. But without the programming previously offered by parks and recreation departments, he explained, facilities that may still be well maintained are now more expensive for people to use. He argued that the loss of public commitment to funding sport and recreation at the programmatic level results in a system oriented toward capitalism, which provides a place to play but lacks the programmatic structure to help youth develop competence and confidence to engage in physical activity for a lifetime.
To build political will for investment in and activation of sport and recreation infrastructure that can reach those most in need, Pathik advocated for a multistakeholder approach. A major medical partner is indispensable to such an approach, he maintained, noting that medical partners are interested in sport and recreation facilities not simply because of sports medicine but because these facilities provide venues for wellness and prevention programming and opportunities to engage with individuals and tailor their products and services. Pathik mentioned parks and recreation directors as another important group of stakeholders. They typically have little influence in the budgeting cycle, he explained, but they are well versed in the programs they offer and can provide metrics to indicate how those programs are being used by youth.
Pathik reported that to equip and empower parks, recreation directors, and advocates to express the value of sport and recreation, the Sports Facilities Companies and the Florida Recreation & Park Association produced calculators that generate reports on how individual parks contribute to property values, health care savings, informed decisions on environmental and public spending, and job creation in the community. The calculators are available for every park in Florida, he noted, and will soon be available for parks in California and Kansas. According to Pathik, this is the kind of
ground-level data that systems science modeling has inspired the organization to try to activate, and it has resulted in billions of dollars of public investment in sport and recreation assets.
Farrey concluded the panel’s presentation with a summary of Project Play’s impact since the 2015 launch of its Sport for All, Play for Life playbook. Based on a 2018 survey of Project Play Summit attendees, more than 100 organizations have taken actions guided by the playbook, he reported. ESPN created an “access to sport” prong in its Corporate Citizenship program and has informed Disney’s strategies, Farrey continued, and cities and counties have begun to design grants and policies based on the playbook’s eight strategies (see Box 3-1). Project Play has helped unlock or shape more than $55 million in foundation and corporate grants, he noted, as community stakeholders have recognized opportunities for investment in youth access to sports and revised their programs accordingly. He highlighted this investment as an outcome of a systems science approach to generating relevant data to share with business leaders and community stakeholders. Farrey also highlighted Project Play’s role in helping shape the U.S. Department of Health and Human Services’ first-ever National Youth Sports Strategy,2 a resource for policy makers and decision makers in youth sports. Lastly, he emphasized recent data indicating that, relative to 1 year ago, more children—in every income category—are playing sports, fewer are physically inactive, and more coaches are being trained (The Aspen Institute, 2020a). He stressed that these accomplishments are the result of the broad network of organizations that have participated in Project Play activities and introduced an array of aligned programs.
PANEL AND AUDIENCE DISCUSSION
Following their presentations, Farrey, Lee, and Pathik addressed a participant’s question about using systems science modeling to help avoid exacerbating disparities in access to sports and shared examples of increasing access to sports in underresourced communities.
Addressing Disparities in Access to Sports
Pathik mentioned that the number of children who receive free or reduced-price lunches is a metric used by the Sports Facilities Companies as a proxy to determine how many children need help accessing opportunities for sport in a given community. The organization begins by assessing whether the community’s budget is adequate to sponsor access to sport and
___________________
2 See the National Youth Sports Strategy at https://health.gov/our-work/physical-activity/national-youth-sports-strategy (accessed November 6, 2020).
recreational activities for that number of children, he explained, noting that this is a starting point. Lee added that while trend and correlational data are useful, the right models can help account for the interaction between people’s individual characteristics and the contextual systems in which they live, which affect the ability to access physical activity.
Farrey emphasized the importance of raising awareness of gaps in access through data and outreach to media, noting that many people otherwise tend to assume that most youth play sports. He highlighted that children from families with annual household incomes of ≤$25,000 play sports at half the rate of children from families with annual household incomes of ≥$100,000—a national metric that Project Play tracks annually (The Aspen Institute, 2019). That metric can be fine-tuned for a specific community, he added, based on the availability of local data.
Project Play’s Work in Underresourced Communities
Farrey recounted the initiative’s work in Harlem, where it surveyed children in schools to ask what sports they wanted to play and learned that hockey was of high interest. He said that Project Play then shared these results with the National Hockey League’s (NHL’s) closest team, the New York Rangers, and venue, Madison Square Garden, and connected them with schools to get floor hockey into many schools in East Harlem. Children in Detroit also expressed interest in hockey, Farrey recalled, but fewer than 1 percent were playing. He reported that the NHL and the Detroit Red Wings, together with other partners, developed a program to introduce floor hockey in schools, which will give tens of thousands of children the opportunity to try the sport.
Pathik shared an example from Rocky Mount, North Carolina, where the Sports Facilities Companies helped open an event center. The town debated about whether to locate the facility in a blighted downtown area or in the suburbs, where it would be near hotels and other destinations, and ultimately used a federally qualified census tract downtown to allow the project to qualify for new market tax credits. A federally qualified medical partner was built in the same area, Pathik added, which opened access to mammography services for women living nearby. He noted that prior systems science modeling data assessing the conditions and needs in that census tract were instrumental in the event center’s placement in an area of high need.
COMMUNITIES: BALTIMORE, MARYLAND
In the second panel, three speakers discussed the use of systems science modeling in communities—specifically in Baltimore, Maryland—offering
another example of coordination and partnership. The speakers, who took turns presenting as they covered different portions of the presentation, included Bruce Y. Lee, who had also spoken in the prior panel; Joel Gittelsohn, professor in the Center for Human Nutrition at the Johns Hopkins Bloomberg School of Public Health; and Sarah Buzogany, food resilience planner for the Baltimore Food Policy Initiative in the Baltimore City Department of Planning.
Lee began by sharing an outline for the presentation, which would begin with an overview of the Baltimore food system, followed by discussion of a specific community project and then three examples of how systems science modeling had informed potential policies and interventions in the city and engaged community and other stakeholders. He reminded participants that obesity is a systems problem that results from the interaction of multiple complex systems of factors, such as biological processes, behaviors, social networks, the environment, policies, economics, culture, and a variety of other factors. According to Lee, focusing on those various systems is important for addressing obesity but is something humans struggle to do because of the difficulty of discerning processes that are occurring beyond those with direct, immediate effects.
Lee briefly described the Global Obesity Prevention Center, previously housed at Johns Hopkins University, and its goal of using different types of systems science approaches to address the obesity epidemic. He stated that the Center consisted of an administrative core, a systems science core, an education and training program, and various domestic and international projects. One of its main initiatives was a multiscale, multicomponent intervention project (led by Gittelsohn) that ran for more than 5 years and addressed obesity in Baltimore City. According to Lee, the project’s goal was to apply different systems science approaches to understand all of the complexities involved in Baltimore City’s food system, as well as to use new systems science approaches such as systems mapping and systems science modeling.
Buzogany then discussed Baltimore’s food system and food policy landscape and highlighted the benefits of the city’s partnership with Johns Hopkins University and the Global Obesity Prevention Center. For the past 10 years, she said, the Baltimore Food Policy Initiative has driven the city’s food policy agenda in partnership with and alongside residents, organizations, and partners such as Johns Hopkins. The relationships between these institutions and the city have been instrumental, she declared, because academic partners have used systems science modeling to help community stakeholders identify needs, understand the potential implementation of strategies and plans, and suggest viable policy solutions.
Buzogany stated that one of Baltimore’s frames is to use food as a catalyst to address health, economic, and environmental disparities in healthy
food priority areas. A “healthy food priority area,” she explained, is a geographic area of the city where residents may face structural barriers to accessing healthy foods, adding that the phrase evolved from the term “food desert.”3 She noted that nearly one-quarter of the city’s population lives in such areas, which are characterized by low healthy food availability scores, income levels at or below 185 percent of the federal poverty level, low vehicle availability among residents, and located at least one-quarter of a mile from a supermarket.
Next, Buzogany shared a map of healthy food priority areas in Baltimore. The map includes locations for hundreds of small grocery, corner, and convenience stores across the city, which she said often lack healthy foods (based on Healthy Food Availability Index scores), as well as public markets, supermarkets, urban agriculture sites, and food assistance sites (e.g., food pantries and distribution sites for summer meals). According to Buzogany, systems maps have helped the Baltimore Food Policy Initiative provide briefs for elected officials and generate neighborhood maps to help residents understand their food landscape. But because the maps are static, cross-sectional tools, she said, the initiative partnered with Gittelsohn to use the maps to explore policy questions such as the effect of a staple foods ordinance on food quality.
Buzogany highlighted the Baltimore Food Policy Initiative’s eight-point healthy food environment strategy: (1) resident-driven processes, (2) corner and convenience stores, (3) supermarkets, (4) public markets, (5) food distribution and small businesses, (6) federal nutrition assistance, (7) urban agriculture, and (8) transportation gaps. By collaborating on systems science modeling efforts, she added, the initiative was able to understand how it could best implement strategies for improving the food environment within the focus areas of its healthy food environment strategy.
Gittelsohn discussed B’More Healthy Communities for Kids, a multilevel program implemented by the Global Obesity Prevention Center to improve access to, demand for, and consumption of healthier foods and beverages in low-income neighborhoods of Baltimore. He focused on the program’s policy working group, which comprised the city’s food policy director; its food resilience planner; and stakeholders in multiple sectors, including the city council, the city health department, city schools, the family league, parks and recreation, wholesale companies, and academia. Gittelsohn reported that the group engaged in a series of planning exercises and activities to consider and develop solutions for improving the city’s food environment, with the goals of building the evidence base to support
___________________
3 More information on “healthy food priority areas” can be found at https://planning.baltimorecity.gov/baltimore-food-policy-initiative/food-environment (accessed November 16, 2020).
healthier food environment policies in Baltimore City and sustaining activities initiated by the program. Gittelsohn noted that policy working group members asked him and his team to develop a systems science simulation model that would provide evidence for a proposed urban agriculture tax credit.
Buzogany observed that the city’s urban agriculture efforts have increasingly focused on achieving equity, particularly racial equity. She shared an excerpt from Baltimore’s 2019 sustainability plan to convey the city’s intention to focus urban agriculture efforts on historically excluded populations: “A city where communities that have been historically excluded from access to land and to fresh, healthy, culturally-appropriate foods are those that benefit most from urban agriculture” (Baltimore Office of Sustainability, 2019, p. 51). The goals of the 2019 sustainability plan have a strong focus on land ownership, Buzogany emphasized, especially for Black farmers and other farmers of color. She observed that policies such as the urban agriculture tax credit were created to help put the city’s vacant land to productive use and to support ownership by urban farmers without burdening them with large property tax bills.
Gittelsohn elaborated on the urban agriculture tax credit, recounting that city Councilman Pete Welch approached the Global Obesity Prevention Center about his proposed bill to provide a 90 percent tax reduction for owners of vacant lots who converted them to urban farms (i.e., the urban tax credit). The Center had an existing systems science agent-based model called the Baltimore Low Income Food Environment (BLIFE), which Gittelsohn said was modified to provide evidence in support of the potential impact of the bill. The BLIFE agent-based model focused on only one area (constituting about 10 percent) of the city, he explained, and modeled children’s after-school food consumption and activity. He added that the model was parameterized with geographic information systems data from that geographic area, and also incorporated data on gender, age, height, weight, and home address for nearly 300 children residing in Baltimore, simulating their walking, physical activity, and food consumption habits. Results from the simulation modeling were provided to the city council as evidence of the potential impact of the urban farm tax credit, he said, and legislation was passed to establish the tax credit in Baltimore City.
Lee emphasized that although the BLIFE model was useful for testing different policies and interventions, expanding it to include all of Baltimore and incorporate more agent characteristics and aspects of the food environment enabled its use to develop larger-scale, agent-based models for any location in the city. He highlighted a second systems science model—an expanded model called VPOP, which Lee had discussed in the session’s first panel—that uses geocoded data to represent key locations, such as households, large workplaces, schools, physical activity locations (e.g., parks,
gyms, recreation centers), and food sources. The more than 1 million computational agents that were placed in the model represent the city’s actual population, Lee emphasized, because the model’s synthetic population was built using census data, and each agent has an age, gender, race/ethnicity, socioeconomic status, and other characteristics that map to those data. Lee reiterated his earlier comments to explain that as the agents move through daily activities and consume and expend energy, modelers can track population health outcomes, such as changes in BMI percentile, over time.
Lee recounted that the Baltimore City Health Department periodically contacted the Global Obesity Prevention Center to share potential policy interventions, such as placing warning labels on sugar-sweetened beverages in different combinations of grocery stores, corner stores, schools, and other settings where the beverages are available. The VPOP model forecasted the potential effect of the labels on purchasing and consumption of sugar-sweetened beverages. Lee reported that according to the model, a warning label would result in people choosing an alternative to sugary beverages 8 percent of the time, based on several studies that examined the efficacy of such labels (Lee et al., 2018). This finding was used as a baseline, he continued, to forecast the impact of having versus not having these labels at three different levels of efficacy—4 percent, 8 percent, and 12 percent—in three major cities (Lee et al., 2018). The model indicated that as the efficacy of warning labels increased, a greater decrease in the prevalence of obesity would occur, effects that were consistent in direction but not in size across the three cities. Lee explained that the heterogeneity of effects across cities resulted from differences in geographic layouts, types of food sources, and combinations of population BMI distributions and health conditions. He stressed that this heterogeneity shows the importance of using systems science models that explicitly represent the population and food environments in the target city of a policy or intervention.
Lee added that the VPOP model was also used to conduct a sensitivity analysis in which modelers varied factors such as warning label efficacy, population literacy rate, and corner store compliance (Lee et al., 2018) (see Figure 3-4). This is another important use of a systems science model, he noted, because it can forecast the outcomes of changing various aspects of a policy or the circumstances in which it is implemented. Sensitivity analyses help stakeholders identify potential drivers of a policy intervention, he explained, and he suggested that understanding such relationships may be even more important than predicting exactly what will happen. He cautioned against viewing systems science modeling as a “crystal ball,” given that exact predictions may not be accurate but can provide a sense of what may be happening.
Next, Buzogany introduced a third example of applying systems science modeling to inform Baltimore community policy making: a systems science
NOTES: SES = socioeconomic status. This figure summarizes the sensitivity analyses used to investigate how varying different unknown factors affects the change in obesity prevalence for Baltimore City. Each factor analyzed is represented by a horizontal bar, where the width of each bar represents the range of impact on obesity prevalence across a range of values for that factor (listed above the bar). Additional lines on the bar represent mean obesity prevalence from intermediate factor values. The dotted line marks 0 percent change in obesity prevalence, or the change in obesity prevalence at the beginning of each scenario.
SOURCES: Presented by Bruce Y. Lee, September 16, 2020; Lee et al., 2018. Figure used with permission from Elsevier.
dynamics model for exploring the effects of a staple foods ordinance. Resident-driven policy is a key tenet of the city’s healthy food environment strategy, she explained, and this tenet manifested in the creation of a group called Resident Food Equity Advisors. The advisers were tasked with co-creating policy with the city planning department and government, and they convened for 2 years to research and develop recommendations for small retailers, such as corner and convenience stores. One of the recommendations, Buzogany continued, was to require corner stores to carry a minimum stock of health-promoting foods. The group partnered with Gittelsohn to develop a systems science dynamics model to simulate the effects of a staple foods ordinance, which she highlighted as an exemplar of government, community, and academia co-contributing to a shared vision.
Gittelsohn elaborated on the systems science dynamics model, which he said enabled the group to simulate the inclusion of different foods and beverages in differing amounts in a potential Baltimore staple foods ordinance and then use the model to recommend modifications to the ordinance. He explained that a first step in the model’s development was to create a flowchart of the different processes involved, such as factors and constructs relating to supply (i.e., food retailers), consumers, and the ordinance itself. Each variable in the flowchart could be changed, Gittelsohn added, and modelers manipulated the variables to simulate four scenarios in which Baltimore corner stores could implement a staple foods ordinance. The scenarios varied with regard to levels of required minimum stock and enforcement.
Gittelsohn reported that one of the model’s outputs was a projection of weekly profits that corner stores could expect to realize if they implemented a staple foods ordinance under each of the four simulated scenarios. Three of the scenarios, he noted, were projected to turn a small yet significant profit, but one, the Supplemental Nutrition Assistance Program depth-of-stock requirements, was found to be burdensome to the point of being unsustainable in the simulation. Thus, he said, this scenario would likely cause implementing stores to go out of business quickly, essentially removing it from policy consideration. Gittelsohn cited other outputs from the modeling simulation as well, such as information about optimal food prices that would achieve the maximum profitability, amounts of foods to order from suppliers in order to optimize sales and profitability and minimize waste, and expected levels of waste and consumer demand.
Gittelsohn briefly listed the model’s limitations, starting with the specificity of the parameterization data to Baltimore City. Because local consumers and retailers provided the data, he elaborated, modification would be necessary to make the model usable in other settings. He pointed out further that the model is focused on proximal effects of a staple foods ordinance and is not yet linked to more distal effects, such as obesity or health outcomes.
Lee summarized the panel presentations by reiterating that the Baltimore case study illustrates how systems science modeling can better address and illuminate the complexities of a community, serve as a virtual community for testing different policies and interventions, facilitate communications and engagement within a community, and bring diverse stakeholders together. He ended by emphasizing the value of collaboration between systems science modeling experts and community decision makers and stakeholders, which he said enables an iterative feedback loop that informs and improves study design, data collection, and modeling parameters and supports other types of research.
PANEL AND AUDIENCE DISCUSSION
Following the second panel’s presentations, Lee, Buzogany, and Gittelsohn commented on how dynamic systems components influence systems science modeling efforts and answered participants’ questions about validation of models, strategies for communities that want to get started with modeling, training and education to prepare decision makers for systems science modeling, and support for urban agriculture tax credit recipients.
The Influence of Dynamic Systems Components on Systems Science Modeling Efforts
Lee explained that one of the reasons for developing an explicit model that represents all of the processes, components, and mechanisms of a system is that changes in policies and priorities may occur. It is helpful for a model to represent what is happening on the ground in terms of the actual mechanisms and processes, he observed, because if policies or interventions change, then modelers can adjust what is happening to the mechanisms. On the other hand, he continued, when a model is highly tailored to a specific situation, changes in outcomes of interest warrant returning to the model to manipulate its inputs.
Gittelsohn referenced the experience of the systems science dynamics model on a staple foods ordinance as an example of an iterative process. The team discussed different expectations and policy needs, he explained, and evolved the type of staple foods ordinance it considered as needs and priorities changed within the city government. Buzogany highlighted the role of political will in enacting, implementing, and enforcing policy. She noted that leadership changes in Baltimore have led the Food Policy Initiative to consider policy options in terms of the degree of political will present to move them, adding that even if a model indicates that a policy will have the intended impact, political will is a critical moderating factor.
Validation of Systems Science Models
Lee responded to a question about how to validate a systems science model of an intervention for reducing the prevalence of obesity. Given that many factors drive the prevalence of obesity, he stressed the importance of validating multiple aspects of the model (i.e., direct and indirect effects) and using more than one measure to validate each aspect. He explained that ground-level data can be used to validate how model agents move about and make decisions, such as by reviewing historical and current food purchasing and consumption data and matching agent behaviors accordingly. Weight changes in a model’s synthetic population can also be programmed, he added, to mirror historical changes in the prevalence of obesity. According to Lee, when modelers validate a model in multiple different ways to help ensure that it represents all that it is supposed to represent, they can feel more confident that the model will take into account all of the important factors and components when a new policy is introduced. He stressed that the validation process never stops because although a model cannot achieve 100 percent validation, it is still important to continue to generate evidence to increase confidence that it is representing what it is intended to represent.
Getting Communities Started with Systems Science Modeling
Gittelsohn discussed the groundwork involved in building relationships within Baltimore, explaining that the Johns Hopkins team has been working in Baltimore communities for at least 15 years toward the goal of improving the food environment and that the team’s modeling efforts began 5 or 6 years ago. Part of the secret of advancing such a goal, he disclosed, is establishing relationships with people in the community, maintaining those relationships over an extended period of time, and demonstrating commitment to working for change. Also helpful, he added, was that the university had already conducted other interventions in the city’s corner stores and carry-out restaurants, which provided data that contributed to the model-building activity. As for funding, he noted that CDC’s Nutrition and Obesity Policy Research and Evaluation Network and the National Institutes of Health have interest in this type of work.
Buzogany highlighted Baltimore’s proximity to Johns Hopkins University, noting that the Food Policy Initiative benefits from its relationship with the university and its access to models developed there. Lee urged participants not to be dissuaded if they are starting from scratch to build relationships in communities. He encouraged them to start at the micro level, such as by examining a small, 1- or 2-mile radius area of the community to assess what is happening. Doing so will keep the effort low-cost
while giving community stakeholders a concrete example of simulation modeling, he elaborated, which could help ease them into investing resources to develop a broader model.
Training and Education to Prepare Decision Makers for Systems Science Modeling
The speakers responded to a question about the extent of training and education required to get started with systems science modeling. After noting significant variability in states’ levels of understanding of and experience with systems science modeling, Lee suggested providing decision makers unfamiliar with modeling concepts with research articles that can help them understand how models can aid in decision making. He emphasized the importance of avoiding providing articles with complex, technical jargon, lest decision makers perceive modeling as an arcane, academic exercise. In addition to sharing research, he suggested simply talking to and discussing ideas with people.
Gittelsohn pointed out that stakeholders may not recognize the iterative, multistep process that occurs between deciding to develop a model and conducting a complex simulation. This is a participatory process, he explained, that includes substantial planning and discussion of concerns, main players, and key factors, and occurs before mathematical calculations and equations come into play. Gittelsohn also suggested emphasizing that the group model-building process involves community members and decision makers and is not exclusive to academics or researchers, so that all types of stakeholders understand the critical role of their participation in the process.
Lee reiterated Gittelsohn’s point that systems science modeling is an iterative process, and added that stakeholders can begin by developing a relatively simple model. There is value in developing an initial, smaller conceptual model to help give people a better sense of the topic of interest, he elaborated, even though this initial model may lack certain data and not provide a perfect representation of every factor involved. Once the initial model has been developed, he continued, stakeholders can prioritize areas where more data are needed and then collect those data to feed into the model, making it increasingly more comprehensive as the cycle continues.
Support for Urban Agriculture Tax Credit Recipients
Buzogany responded to a question about how recipients of the urban agriculture tax credit were prepared to succeed in converting their land for agriculture. She cited three organizations that offer such support: Future Harvest, which has a beginner farmer training program that pairs
novice farmers with existing farms to facilitate hands-on learning; the Farm Alliance of Baltimore City, a member organization (which includes more than 20 of the city’s urban farms) that shares tools and expertise and advocates for policy change; and the Black Church Food Security Network, which organizes beginner farmer training through Black churches and networks.
POLICY MAKERS: SALURBAL
The third panel included three speakers who discussed the usefulness of systems science approaches for policy makers. The speakers shared experiences with policy maker engagement in an international collaborative initiative called SALURBAL (Salud Urbana en América Latina [Urban Health in Latin America]), which incorporates systems thinking and systems approaches to promote urban health and environmental sustainability in Latin American cities.
Ana Diez Roux, dean and distinguished university professor of epidemiology in the Drexel University Dornsife School of Public Health, opened the panel with an overview of the SALURBAL study’s goals and aims. It is an initiative of the Urban Health Network for Latin America and the Caribbean, she explained, implemented by Drexel University and 14 partners based primarily in Latin America and funded through the Wellcome Trust’s “Our Planet, Our Health” global initiative. Diez Roux cited the study’s three goals: to create an evidence base for making Latin American (and other) cities healthier, more equitable, and environmentally sustainable; to engage policy makers and the public in a new dialogue about urban health and urban sustainability and implications for societal action; and to create a platform and network that will ensure continued learning and translation. She also described the study’s four aims: to identify city and neighborhood drivers of health and health inequities among and within cities, based on an analysis of data pooled from various sources; to evaluate the health, environmental, and equity impacts of policies and interventions by capitalizing on natural experiments; to use systems thinking and simulation models to evaluate urban–health–environment links and plausible policy impacts; and to engage the scientific community, the public, and policy makers in disseminating and translating the study findings.
Diez Roux highlighted SALURBAL’s systems thinking aim, noting that systems science approaches are integrated into the full context of the study as one of three methods (alongside observation and experimental evaluation) for obtaining evidence relevant to urban health. She explained that the researchers decided to apply systems science modeling to transportation policy and food policy, which were selected as the two areas of emphasis because feedbacks and dependencies (in the sense of people influencing
each other through social norms) are critical to both. She added that both areas have high policy relevance globally and regionally, and have strong health–environment links that may operate in both directions, which means that the effects of policies will likely involve reinforcing or buffering loops or dependencies.
Diez Roux emphasized the uniqueness of the policy maker engagement component of the study’s fourth aim, suggesting that its purpose is to ensure the study’s policy relevance, as well as to promote new ways of thinking among policy actors and other stakeholders about the drivers of urban health and the types of policies and interventions that could improve health and sustainability in cities. She explained that the study has engaged policy makers in its systems science aim by including them in participatory group model-building activities conducted in three regional workshops, which she said simulated local applications of this participatory approach. The study will conduct a second policy maker engagement activity once development of two simulation models has been completed, she added, and then yielded to her fellow panelists to elaborate on these examples of policy maker engagement.
Brent Langellier, assistant professor of health management and policy at Drexel University, expounded on SALURBAL’s participatory group model-building exercise and its two policy-oriented, agent-based models. He began by reiterating the study’s systems thinking aim:
To employ systems thinking and formal systems simulation models to (1) better understand the dynamic relations between the urban environment, health and environmental sustainability; and (2) identify the plausible impacts of selected policies under varying conditions and dynamic relations.
The terms “models” and “modeling” are used broadly in the study, he clarified, to describe its efforts to both map the systems that influence urban health and employ formal quantitative simulation models, which together can enhance understanding of potentially highly effective policy opportunities.
Langellier stressed that policy makers generally recognize the complexity of urban systems and accept this attribute of their cities. He outlined three policy-relevant aspects of this complexity: feedback loops (bidirectional relationships with two or more variables), interdependence (outcomes of one person are often not independent of outcomes of other people), and change over time (a policy’s effect may depend on the state of a system). He encouraged the use of common tools and terminology to describe complex systems and explore their influences on health, noting that the first 2 years of the SALURBAL study were strongly focused on this objective.
Langellier elaborated on three group model-building workshops—one in Lima, Peru; one in São Paulo, Brazil; and one in Antigua, Guatemala—conducted with 62 policy stakeholders during the study’s first 2 years. These stakeholders included academics, elected and administrative policy makers in spaces such as public health and planning agencies, and representatives from nongovernmental organizations that influence local policies. The outcomes of interest were healthy eating and mobility/transport in Latin American cities, Langellier continued, and the objectives were to engage policy stakeholders in SALURBAL, introduce systems thinking by explaining such concepts as feedback loops and describing how to depict complex systems, describe the structure and function of complex systems that drive both healthy eating and transport and explore their overlap, explore multisector influences on urban health, and identify policy solutions that recognize complex system influences on health outcomes.
According to Langellier, a critical component of the workshops was scripted activities to introduce systems thinking and work toward mapping systems by facilitating the transfer of participants’ mental models to a more explicit form. The workshops began with an activity called “graphs over time,” he recounted, which was intended to engage participants in framing a problem, initiating mapping, generating variables, and prioritizing the variables as preparation for creating causal loop diagrams. Participants were prompted to think of a factor that influences healthy eating in cities, and to draw one trajectory showing how they hoped that factor would change over time and another showing how they feared it would change over time.
Another scripted activity was the creation of causal loop diagrams, Langellier continued, in which participants were instructed to create such diagrams in small groups and then present them to the larger group. A synthesis activity followed in which the small groups worked with facilitators to synthesize their diagrams into a single causal loop diagram representing a hypothesis about the food system/transportation system factors that influence a healthy urban environment. This juncture, Langellier said, is where the diverse and often overlapping perspectives of stakeholders emerged as the synthesis model took shape, and participants quickly recognized the complexity and interconnectedness of the systems involved. Langellier shared an example of an initial causal loop diagram produced during one of the workshops (see Figure 3-5). He flagged feedback loops related to healthy food consumption in the bottom left of the graphic and compared them with feedback loops related to transport and mobility incidents, located in the top right of the graphic. Stakeholders naturally integrated these two realms without prompting from the facilitators, he recalled, which he said illustrated the concept of multisector influences.
Langellier explained that after the causal loop diagrams had been created, participants completed a final scripted activity called “action ideas,”
NOTE: AQ = air quality; PA = physical activity; UPF = ultraprocessed food.
SOURCES: Presented by Brent Langellier, September 16, 2020; Langellier et al., 2019.
which he described as an opportunity to reflect on the system’s structure and feedback loops and consider how they may influence the desired and undesired changes over time, helping to generate ideas for changes to the system to improve health outcomes. Those changes could take many forms, he clarified, such as adding or removing variables or feedback loops, or changing their magnitudes. After stakeholders had devised policy interventions, they added them to a 2 × 2 grid that ranked them by ease of implementation (easy or hard), as well as by level of impact (high or low). This exercise, Langellier observed, helped them consolidate the easy-to-implement, high-impact ideas, which in the workshops included advocating for food and transport, promoting cycling tours between healthy food sources in the city, and offering microcredit and credit assistance programs for urban farming.
Langellier shifted to a review of SALURBAL’s two agent-based simulation models of urban policy, one focused on ultraprocessed food purchasing policy and the other on urban transport and mobility policy. He prefaced this discussion by noting that the stakeholder insights and action ideas generated during the group model-building workshops were important
inputs into the formal model development, along with the research team’s knowledge and literature reviews.
Langellier began with the ultraprocessed food purchasing model, explaining that it explored two questions. The first was how food labeling, advertising, and taxes can be combined most effectively to reduce purchasing of ultraprocessed food in Latin American cities. Langellier observed that there is currently much policy action on taxation, labeling, and advertising policies in numerous Latin American cities and countries. He characterized this as an opportunity to “strike while the iron is hot.” He then shared the second question: whether policy effects vary across different population segments (high versus low income and educational attainment). Langellier explained that the model comprises adult females living in urban areas in Latin America who are the primary food purchasers in their households, and uses population demographics and ultraprocessed food purchasing data from Mexico City. He noted that those data could easily be replaced with data from any other urban area in Latin America, providing a flexible framework. The agents are organized in social networks, he continued, with ties more likely among agents of similar age, income, and education levels. He added that, because social influences and social norms are important drivers of dietary behaviors and are potentially important for policies aimed at reducing inequities, it was important for the model to account for inter-dependencies among people.
Langellier pointed out that as the model runs, ultraprocessed food purchasing is updated in response to social norms and social influence and to price, labeling, and advertising policies. He described three factors on which policy effects are based: price elasticities from Latin American countries (price changes relative to income), advertising (advertising elasticities), and labeling (based on a Chilean policy evaluation of the first comprehensive stoplight labeling policies in Latin America).
Moving from the structure of the model to its utility, Langellier reiterated that taxation and labeling policies are under consideration in several Latin American countries, and that policy effects are likely to vary with different contexts, levels, and combinations. He suggested that the food industry will probably respond to such policies with advertising, referencing a history of increased industry advertising and lobbying efforts to influence policy enactment and implementation. An increase in advertising can challenge efforts to evaluate the effects of a policy, he explained, because the change in advertising occurs concurrently with the policy change. Finally, Langellier stressed that developing the modeling infrastructure was helpful because modelers can quickly change the model’s parameters—related to both the city and the tax or policy levels—making the model what he characterized as an urban food policy lab that can be applied to future policy proposals in different locations.
Langellier moved on to discuss the second agent-based model, which deals with urban transport and mobility policy. He described its purpose as modeling commuter decision making and behavior in a city inspired by Bogotá, Colombia, where there are five commuting modes: car, motorbike, bus and bus rapid transit (BRT), bicycle, and walking. He explained that because SALURBAL’s intent is to explore interconnections among behavior, health, and the environment, this model can be used to explore how public transportation policies, taxation, and interventions to improve personal safety from crime affect mode share, physical activity, and air pollution.
The model is spatially explicit, Langellier pointed out, which accounts for residential segregation in a city, income inequity, and commutes between homes and workplaces. Additional features of the model’s design include variance in car, motorcycle, and bicycle ownership by income; a realistic distribution of bus and BRT stops through the city; assignment of a safety variable to each mode based on available crime data; and population and environment factors informed by Bogotá data. Langellier also listed the rules for the urban transport and mobility policy model. Agents make a daily commute-to-work decision, he said, based on the modes of transportation available to them, their determination of the highest-utility mode (a function of travel time and cost), and their evaluation of the perceived safety of each mode (i.e., avoiding modes that they consider unsafe based on their threshold for safety).
Finally, Langellier echoed his earlier comment about the models’ value in establishing modeling infrastructure to enable examination of specific policy proposals. He underscored that both models effectively integrate health aspects with outcomes of interest to policy makers in multiple sectors.
Felipe Montes Jimenez, associate professor in the Department of Industrial Engineering at the Universidad de Los Andes, shared a case study of an effort called TransMiCable that implemented cable car transport in a low-income, densely populated area of Bogotá. Implemented in 2018, the cable car system had more than 7.5 million passengers in its first year of operation. Montes Jimenez was involved in research designed to provide evidence for an urban development effort that implemented 16 projects aimed, for example, at improving facilities for recreational and cultural activities and making physical improvements to homes. He explained that the objective of the research was to evaluate the effect of the implementation of one of these projects—TransMiCable—in three areas: environmental and social determinants of health, physical activity (for both leisure and transport), and health outcomes (quality of life, respiratory diseases, and homicides). A key component of the evaluation, he elaborated, was its use of citizen science to identify, prioritize, and communicate the TransMiCable cable car intervention’s negative and positive impacts on health and quality of life in
the community, which Montes Jimenez said contributed to empowering the community and increasing its uptake of the intervention.
Montes Jimenez then described the evaluation’s conceptual framework, explaining that the first step was to convene multisector stakeholders, including community members and academics. The stakeholders participated in a group model-building workshop where they engaged in scripted activities for the purpose of developing causal loop diagrams, he elaborated, that would complement the conceptual model and identify and explore policy alternatives within the TransMiCable evaluation. He reported that the first draft of the causal loop diagram seemed messy, but that concrete constructs and cause-and-effect relationships were deduced on closer examination and additional dialogue with participants. Montes Jimenez relayed that after the workshop, researchers held one-on-one interviews with several stakeholders to review the diagram, collect inputs on its feedback loops, and clarify terminology to accurately reflect the jargon of each discipline and sector. Following these interviews, the researchers revised the diagram so that it more clearly depicted four key aspects of the system (see Figure 3-6).
Montes Jimenez elaborated on these four aspects, beginning with health conditions. The updated diagram, he said, conveyed that social dynamics favor both physical activity promotion and the dynamics of transport that could influence mental health and leisure activities. He cited as a second aspect social and economic development, denoting social dynamics that could affect mobility, interurban displacement, and quality of life. Mobile phone apps track movements in a sample of people, he added, which helps researchers understand how the intervention is changing mobility in terms of travel time and quality. The third aspect, operationalization of TransMiCable, refers to the provision of an efficient service. Finally, Montes Jimenez described the fourth aspect, citizen’s culture, as involving community participation and ownership of the intervention, as well as reinforcement of inclusive behaviors that nurture well-being and reduce vandalism and gender-based violence in the area where the cable cars operate.
Montes Jimenez ended his presentation with reflections on the TransMiCable evaluation. The involvement of multisector stakeholders led to a shared mental model of the system, he observed, represented by four domains that were previously undefined. He stressed that inviting policy makers to co-create the conceptual framework for research and project evaluation helped make them supportive of the evaluation’s conduct and made its results visible for the city. He noted that a second cableway is intended to be added in the control area of the city as part of a second intervention. Lastly, he highlighted how the involvement of both policy makers and community members served to level the playing field by giving everyone the same opportunities to be heard and to contribute to a shared mental model that incorporated different points of view.
SOURCES: Presented by Felipe Montes Jimenez, September 16, 2020; Sarmiento et al., 2020.
PANEL AND AUDIENCE DISCUSSION
Following the third panel’s presentations, Diez Roux, Langellier, and Montes Jimenez commented on the origins of the Urban Health Network, modeling’s role in informing policy making, modeling training for the public health workforce, and citizen science as a means of engaging communities in systems approaches. They also responded to participants’ questions about SALURBAL’s use of agent-based models versus system dynamics models to simulate intervention impacts, stakeholder feedback from group model-building workshops, and managing strong personalities in these workshops.
Origins of the Urban Health Network
According to Diez Roux, the Urban Health Network originated at Drexel University with a meeting among a handful of colleagues with personal connections to Latin America and shared research interests in urban health issues. The policy maker engagement piece was highly important from the start, she recalled, clarifying that the group’s focus is not on research but on policy engagement. As the Urban Health Network grew and added members in different Latin American countries, she continued, an opportunity arose to apply for funding from the Wellcome Trust, which was interested in funding an international collaboration that involved policy maker engagement. That funding was critical, Diez Roux noted, for creating the Network’s infrastructure and a platform for collaborating and generating evidence to support public policy. She remarked that a similar approach could work for connecting stakeholders in various U.S. metropolitan areas who have shared interests.
Modeling’s Role in Informing Policy Making
Using COVID-19 models as an example, Langellier emphasized the importance of recognizing that although policy decisions do not always match what models may suggest should be done, the models are still influential in generating important insights that influence policy directly or indirectly (e.g., in the case of COVID-19 models by helping to clarify key aspects of the pandemic and how it evolves). COVID-19 models could be developed relatively quickly, he pointed out, because of decades of prior research and investment in the infrastructure needed to develop infectious disease models. In contrast, he pointed to chronic diseases, for which there is a long way to go to build modeling infrastructure that can be responsive to acute policy needs. He explained that a focus of the Urban Health Network’s efforts to develop agent-based and group-based models is to enable
the generation of policy-relevant evidence, but the Network is also building a modeling infrastructure that will facilitate faster responses to future policy decisions.
Modeling Training for the Public Health Workforce
Langellier mentioned a training program in systems thinking and a short course on group model building that were both offered at Drexel, noting that the latter was attended mainly by people from nongovernmental and government organizations who wanted to learn how to use these tools. He highlighted the critical importance of the Urban Health Network’s funding support, which he said allowed it to prioritize the development of modeling infrastructure and capacity. Yi concurred and suggested that funding can also facilitate opportunities to convene and connect stakeholders for informal consensus building and discussion of shared interests and goals.
Community Engagement Through Citizen Science
Montes Jimenez commented on TransMiCable’s citizen science component, whereby community members were invited to help researchers gather data by taking pictures of the environment surrounding the cable car location and observing barriers and facilitators for health. The purpose of the data gathering was to provide useful insights for policy makers, he explained, and ultimately to make agreements and commitments with them about follow-up actions for both the community members and the policy makers. Inviting the community members and policy makers to join in the modeling process made them aware of the initiative and helped them feel engaged in it, he continued, making it more inclusive and systemic because all of the actors participated and contributed to a shared outcome. He added that a citizen science model can foster longer-term sustainability of an intervention because it enables community ownership of an effort.
Use of Agent-Based Models to Simulate Intervention Impact
Langellier responded to a participant’s question about the use of agent-based versus system dynamics models to simulate impacts of the SALURBAL intervention. He noted that SALURBAL collaborators are in the process of developing system dynamics models of obesity transitions in Latin American countries, but explained that agent-based models were chosen for the ultraprocessed food model for purposes of flexibility and for utility in addressing future research questions. The model will examine ultraprocessed food consumption as an outcome, he reminded participants, and although access to different types of food is an important factor, the
model is not spatially explicit. But the agent-based framework will enable researchers to examine how the agents’ ultraprocessed food purchasing behaviors are affected by social networks and social influence, he explained, which are drivers of dietary behaviors and thus are important to consider when developing policies to address social inequities. He pointed out that the model could be expanded in the future to a more choice-oriented framework whereby agents would choose between ultraprocessed and other types of foods.
Stakeholder Feedback from Group Model-Building Workshops
Diez Roux clarified that a range of health care providers, such as physicians and nutritionists, are part of the SALURBAL community, and remarked that the diversity of backgrounds among participants has enhanced the initiative’s group model-building exercises. According to Diez Roux, feedback from these exercises indicated that participants had begun to apply systems thinking to their conceptualization of problems, and that they went away from the exercises with the intention of conducting similar activities at their own worksites. She asserted that this shift in participants’ mindsets is extremely valuable, suggesting that this outcome may be even more valuable than the results generated from the formal simulation models.
Managing Strong Personalities in Group Model-Building Workshops
Montes Jimenez responded to a participant’s question about how to prevent outspoken, opinionated participants in group model-building workshops from suppressing other participants’ involvement in the activity. He explained that facilitators of group model-building workshops are trained to anticipate such situations, to which they may respond by ensuring that all participants have an opportunity to talk and provide feedback on ideas, or by placing participants in small, diverse groups. The goal of developing a shared mental model helps, he added, because it removes the focus from any individual’s specific situation and views.
This page intentionally left blank.
