School Disease Spread Policy Models

ISEF Category: Biomedical and Health Sciences

Subcategory: Other  ·  Difficulty: Advanced  ·  Setup: Home Setup  ·  Time: Full Year

The Hook

One crowded hallway can change an outbreak curve fast. A school is not one big room, it is a web of class periods, lunch tables, buses, and clubs. That makes it a great place to test which policy cuts spread without wasting money. Your model can turn those choices into numbers.

What Is It?

Think of this project like a tiny digital school. Each student, teacher, and bus rider follows simple rules, like who they meet, how long they stay close, and whether they wear a mask. The model then tracks how an infection moves through that network, one contact at a time.

Calibration means you tune the model until it matches real outbreak data from schools or public health reports. Once the baseline looks close to reality, you can change one rule at a time, like adding better ventilation, stronger mask use, or tighter cohorts. That lets you compare policies in a controlled way, instead of guessing which one works best.

Cost-optimal means the best outcome for the money. You are not just asking which policy cuts the most cases. You are asking which policy gives the biggest drop in spread per dollar, which is the kind of question schools actually face.

Why This Is a Good Topic

This is a strong science fair topic because you can test clear policy changes, compare them against a baseline, and turn a messy public health question into measurable results. It connects to real decisions about HVAC upgrades, mask rules, and scheduling. You can learn modeling, data cleaning, calibration, and cost-effectiveness analysis without needing a wet lab.

Research Questions

• How does classroom ventilation change the peak number of infectious students in a simulated school?
• What is the effect of mask compliance on outbreak size when the contact network stays fixed?
• Does cohorting reduce total infections more than changing lunch and extracurricular mixing patterns?
• To what extent do flu, norovirus, and RSV differ in the policy mix that minimizes cases per dollar?
• Which calibration choices best match public outbreak datasets, daily cases, attack rate, or absenteeism?
• What is the effect of class size on the intervention cost needed to keep the outbreak below a set threshold?

Basic Materials

• Laptop or desktop computer with at least 8 GB of RAM.
• Spreadsheet software such as Google Sheets or LibreOffice Calc.
• Python with Jupyter Notebook.
• Public outbreak datasets from CDC FluView, CDC RSV reports, or published school outbreak papers.
• A lab notebook or notes app for assumptions, parameter choices, and model outputs.

Advanced Materials

• A faster computer or cloud notebook environment for running many simulation trials.
• Python with NumPy, Pandas, SciPy, and Matplotlib.
• NetLogo for quick agent-based prototypes.
• R for sensitivity analysis, uncertainty intervals, and statistical tests.
• An anonymized school contact or schedule dataset, if your school or district approves access.

Software & Tools

• Python: Builds the simulator, fits parameters, and plots the results.
• Jupyter Notebook: Keeps code, notes, and figures together while you iterate.
• NetLogo: Lets you prototype agent-based rules and watch spread on a school grid.
• R: Helps you run sensitivity analysis, compare scenarios, and graph uncertainty.
• LibreOffice Calc: Cleans small datasets and checks your summary tables without paid software.

Experiment Steps

1. Define the school contact structure you want to simulate, including classes, lunch, buses, and clubs.
2. Choose the outcome metric you will compare, such as total cases, peak cases, absenteeism, or cost per case averted.
3. Calibrate the baseline model to one public outbreak dataset before you test any policy changes.
4. Build separate scenarios for ventilation, masking, and cohorting so you can compare each policy against the same starting point.
5. Add uncertainty ranges for transmission, compliance, and reporting so your final ranking has error bars.
6. Test cost-effectiveness across several school sizes or disease types to see which intervention wins under different conditions.

Common Pitfalls

• Fitting the model to one outbreak and then judging the same outbreak again, which makes the calibration look stronger than it is.
• Treating all classroom contacts as equal, which hides the extra spread from lunch, buses, and sports.
• Testing masks, ventilation, and cohorting one at a time without a baseline run, which makes the comparison unfair.
• Ignoring behavior changes like missed classes or uneven mask use, which can swing the results.
• Ranking interventions by cases only, which misses cost and can favor an option that is too expensive to scale.

What Makes This Competitive

A class-level project becomes much stronger when you validate on one dataset and test on another, instead of tuning and scoring on the same data. It also helps if you compare several diseases, not just one, because flu, norovirus, and RSV spread in different ways. Add uncertainty bounds, a clear cost model, and a sensitivity analysis, and you get a result that says more than just which policy won in one setup. That kind of careful design makes the work feel real, not just simulated.

Project Variations

• Model flu spread in middle school versus high school to see how age structure changes the best policy.
• Compare ventilation upgrades against mask rules using the same cost-effectiveness metric.
• Test how lunchroom mixing, bus rides, or sports teams change outbreak size under the same school rules.

Learn More

• CDC FluView: Weekly U.S. influenza surveillance data and reports on the CDC website.
• CDC RSV-NET: RSV hospitalization surveillance data and reports on the CDC website.
• NIH PubMed: Search review articles on school transmission, ventilation, masks, and agent-based disease models.
• NCBI Bookshelf: Free chapters and textbooks on epidemiology and infectious disease modeling.
• NetLogo Models Library: Example agent-based models and documentation on the NetLogo website.
• PLOS ONE: Open-access papers on disease modeling and intervention comparisons, searchable on the journal site.