Designing an Agent-Based Model for Childhood Obesity Interventions: A Case Study of ChildObesity180
TOOLS AND TECHNIQUES — Volume 13 — January 7, 2016
This flowchart illustrates the process by which agents update their properties. Body mass index is updated daily based on the agents’ weight and height (weight in kilograms divided by height in meters squared). Working backwards, height is updated based on the growth charts published by the Centers for Disease Control and Prevention, conditional on age and sex, and weight changes based on calorie surplus (7,700 calories per kg). Calorie surplus, in turn, is determined by the difference between daily caloric intake and caloric expenditure. Caloric expenditure is the product of resting metabolic rate and a physical activity multiplier, and resting metabolic rate is computed using the Schofield equations (a function of age, sex, and height).
Figure 1. A visual representation of the agent-based model describing BMI dynamics in each agent. Abbreviations: BMI, body mass index; PA, physical activity; RMR, resting metabolic rate.
Figure caption explains image in detail.
Figure 2. Overview of the agent-based modeling Netlogo interface. Each component of the model was programmed by the modeling team. Fields in the left column indicate each intervention or intervention component. For Town-Type (top of left column), users can select one of 3 town types (town A, town B, or town C); here town type A is selected. Moving down the column, for the Active Schools Acceleration Project (ASAP), each of 3 programs (Program 1, Program 2, and Program 3) is represented by an on/off switch; each program can be turned on or off independently of one another. For Healthy Kids Out of School (HKOS), each of 3 programs (Drink Right, Move More, and Snack Smart) is represented by an on/off switch; each program could be turned on or off independently of one another, but generally HKOS is treated as 1 intervention with all components turned on. The setup button initializes the simulation, creating agents according to assigned properties. The “go” button instructs the agents to carry out their behaviors. Two fields display outputs for the day and time reached by the simulation. A sliding scale for after-school participation characterizes the proportion of children who participate in the after-school program. At the bottom of the column, 2 fields show outputs of the percentage of children who are overweight and obese, by sex. Users select the “setup” then “go” buttons to allow agents to move, eat, and exercise in real time (illustrated in the center screen). The right column displays changes in agent or town properties (mean BMI, mean caloric intake, mean daily energy expenditure, and average calorie surplus) over time. Abbreviations: BMI, body mass index; DEE, daily energy expenditure.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, the Centers for Disease Control and Prevention, or the authors’ affiliated institutions.