Recipient Organization
Case Western Reserve University
10900 Euclid Avenue
CLEVELAND,OH 44106
Performing Department
(N/A)
Non Technical Summary
This project will increase children's fruit and vegetable consumption and reduce food waste by developing, disseminating, and evaluating an agent-based model (ABM) simulation tool that identifies systems and behavioral nudge strategies to optimize consumption of fruits and vegetables in school nutrition programs. Most children do not eat the recommended amount of vegetables, and food waste is also an important environmental issue. Currently, behavioral nudge approaches, which aim to reduce waste while improving selection and consumption of vegetables, fruit, and other healthy foods, dominate school nutrition research and practice. Yet, the complexity of school meal environments requires a systems approach that is missing from the available evidence on behavioral nudges. Our overall goal is to empower K-12 school nutrition staff with evidence-based strategies specific to their institution to improve child fruit and vegetable consumption while decreasing food waste through the development and evaluation of a school meal systems ABM. We will use our preliminary data to develop the ABM and test its predictive validity by conducting real-world, nudge experiments in diverse school settings and comparing the experimental findings to the ABM predictions. We will disseminate the ABM nationwide and evaluate the tool's effectiveness among school nutrition professionals. Lastly, we will provide hands-on opportunities for college students to participate in this project and offer a seminar course on using systems approaches to prevent chronic disease. Our ABM tool will fill a critical gap by identifying data-driven strategies to improve child diet quality, mitigate food waste, and reduce school nutrition program costs.
Animal Health Component
25%
Research Effort Categories
Basic
0%
Applied
25%
Developmental
75%
Goals / Objectives
Our overall goal is to empower K-12 school nutrition directors with evidence-based strategies specific to their institution to improve student fruit and vegetable consumption while decreasing food waste through the development and evaluation of a school meal systems ABM simulation tool. This goal will be achieved through four objectives:Obj 1: Build an ABM to simulate dietary behavior, particularly fruit and vegetable selection and consumption, in school meal systems using our preliminary data and peer-reviewed literature. Function: ResearchObj 2: Validate our ABM simulation tool using data from new school nutrition intervention experiments in collaboration with key stakeholders. Function: Extension and ResearchObj 3: Disseminate the ABM simulation tool nationwide and evaluate the ABMs effectiveness in improving the selection, consumption, and waste of fruit and vegetables in K-12 school meals. Function: ExtensionObj 4: To promote research skills and systems approaches among the next generation of nutrition scientists and practitioners, we will involve undergraduate and graduate students in school nutrition intervention implementation, data collection, and data analysis through hands-on research experiences and a transdisciplinary seminar course on using systems approaches to prevent obesity and chronic disease. Function: Education
Project Methods
Obj 1: Build an ABM to simulate dietary behavior, particularly fruit and vegetable selection and consumption, in school meal systems using our preliminary data and peer-reviewed literature. Function: Research1a. Abstract ABM parameters. Parameters are the constants in the equations and algorithms that are used to create the ABM schedule or processes. We will use our preliminary data to assign baseline and trend characteristics to the ABM in the initial ABM parameterization. We will also use publically available administrative data to establish the distribution of student enrollment across age, sex, race, and ethnicity of K-12 students, and we will utilize peer-reviewed literature to supplement these findings.1b. Creation of basic ABM structure in NetLogo. Daily agent schedules will determine the processes in which agents interact with each other in the simulation. General school-level conditions, like student enrollment and time available for lunch, will be incorporated into the user via point and click sliders and radio dials.1c. Calibrate the ABM.Calibration entails establishing appropriate values for key parameters by testing which parameter values cause the ABM to reproduce patterns seen in preliminary data.1d. Verification processes. This refers to the process of determining that the computational ABM accurately reflects the underlying economic and behavioral theory and other important system factors, like relevant school nutrition policies.Obj 2: Validate our ABM simulation tool using data from new school nutrition intervention experiments in collaboration with key stakeholders. Function: Extension and Research2a. Face validity testing. We will present our ABM simulation tool in an interactive meeting to get input from school nutrition experts to determine any flaws in key assumptions or ABM relationships.During the presentation, we will seek feedback on the agent schedules used to develop the ABM, as well as the appearance and usability of the point and click interface. We will revise the ABM simulation tool according to stakeholder comments.2b. ABM Validation and Verification. Using a predictive validation technique, we will conduct 1-year intervention experiments at 3 schools to determine if our ABM simulation tool can predict the meal participation, selection, consumption, and waste results of the new real-world intervention experiments as determined by production records and plate waste assessments. System variables will be collected during each plate waste assessment, such as the length of the lunch period, amount of time for the lunch line to dissipate, recess availability, and weather. Experiment data will be analyzed using Stata version 15 (College Station, TX). Quantitative analysis will consist of repeated measures analysis of variance (ANOVA) for each measurement point for each of the primary outcome variables (fruit and vegetable selection, consumption, and waste). Covariates will also be added to the ANOVA design to explore the effects of school size, lunch period duration, free and reduced-price eligibility, and school-level student race, ethnicity and sex demographics.The simulation settings will be fixed to reflect the system characteristics of the intervention schools (time available for lunch, free/reduced lunch percentage, etc.). The agreement between the real-world and simulated values for participation, selection, consumption, and waste will be assessed using correlation coefficients, t-tests, and two-sample variance-comparison tests. We will compare the observed and simulated values for participation, selection, consumption, and waste. If there are significant differences (p<0.05) between the observed and simulated values, the verification and calibration procedures will be repeated to identify and address the issue(s).2c. Additional face validity testing. We will convene a virtual meeting with School Nutrition stakeholders using Zoom to present the final ABM simulation tool and get an additional round of feedback (and subsequent revision) of the final ABM simulation tool. In particular, these field experts will be consulted to determine if the final ABM simulation tool's input-output relationships seem plausible.Obj 3: Disseminate the ABM simulation tool nationwide and evaluate the ABMs effectiveness in improving the selection, consumption, and waste of fruit and vegetables in K-12 school meals. Function: Extension3a. Simulation Tool Dissemination (i.e. Extension). The ABM simulation tool will be primarily disseminated via presentation at the School Nutrition Association annual meeting, the Academy of Nutrition & Dietetics' annual Food, Nutrition & Conference Expo, and a free online webinar.3b. Extension evaluation. As a part of the ABM simulation tool download process, users will be asked to provide demographic information about their school or district, as well as an email address. We will use the email addresses to distribute a Qualtrics survey approximately 3 months and 6 months after the tool is downloaded. The survey items will include information on why the ABM simulation tool was downloaded, how it was used, suggestions for needed tool improvements, and how the tool has benefited or otherwise impacted their school nutrition program.Descriptive statistics for closed-ended questions and qualitative content analysis for open-ended questions will be used to analyze the survey results. The findings will be used to estimate the reach and impact of the ABM simulation tool, as well as identify opportunities to improve the tool and enhance nationwide generalizability.Obj 4: To promote research skills and systems approaches among the next generation of nutrition scientists and practitioners, we will involve undergraduate and graduate students in school nutrition intervention implementation, data collection, and data analysis through hands-on research experiences and a transdisciplinary seminar course on using systems approaches to prevent obesity and chronic disease. Function: Education4a. Hands-on Research Experiences. Undergraduate and graduate students will gain direct experience in implementing the intervention experiments in Objective 2 (assist Extension and school staff in intervention delivery and collecting intervention fidelity measures), data collection (structured cafeteria and kitchen observations, plate waste assessments), data entry, assisting with plate waste analyses, and documenting feedback from key stakeholders during the face validity processes).4b. Transdisciplinary seminar course. PD Prescott will also teach a transdisciplinary special topics course on systems approaches to preventing obesity and chronic disease open to upper-level undergraduate students as well as graduate students.4c. Education Evaluation. This objective will be primarily evaluated by the number of undergraduate and graduate students who complete hands-on training opportunities and/or the seminar course. In addition, student evaluation scores of the seminar course will be used. Students will use a Likert scale to assess PD Prescott's teaching effectiveness and the overall quality of the course.