A Retrospective Evaluation of Pandemic Policy Impact on University Campus: An Agent-based Modeling Approach for Mobility, Disease Propagation, and Testing During COVID-19

In the fight against the spread of infectious disease, university campuses present a unique set of obstacles because of the prevalence of communal living arrangements and the difficulties of restricting sociability and group gatherings. In the last few years, the dynamics of the pandemic have evolved, and so too have the challenges university campuses face in ensuring their communities’ health and safety. This study introduces a complex and adaptable hybrid model, which was developed, applied, and validated using data from the peak incidents of SARS-CoV-2 during 2020–2021. It combines an agent-based model (ABM) with a modified SEIR system dynamics approach. Recognizing the shifting landscape of the pandemic, this model has been designed to function as both a historical case study of COVID-19 dynamics on a university campus and a framework for addressing future pandemic threats. By simulating disease propagation considering contact distance (i.e., 0–3 feet, 3–6 feet), contact duration (e.g., indoor layout, capacity, ventilation), and individual behaviors (e.g., partying, gathering, sporting, off-campus activities), the model serves as a comprehensive analysis tool. It incorporates GIS-based campus data, real-time Wi-Fi occupancy data, student schedule-based behavioral patterns to closely emulate campus life. The testing frequency (e.g., mandatory, voluntary), methods (e.g., PCR, antigen, antibody, saline gargle, saliva), and different containment policies (e.g., mask-wearing, vaccination) enhance the model's predictive capabilities. The model's flexible structure allows stakeholders to adapt it to current and emerging scenarios. Retrospective analyses indicate that strategies like indoor mask mandates, frequent testing, and high vaccination rates were pivotal in managing spread of disease. The model's predictive accuracy, as evidenced by high accurate rate (i.e., 85.1% accuracy with an average deviation of 4.42 cases per day) when comparing the model output to actual campus data, underscores its value as a decision-making aid for university administrators in the ongoing efforts to foster a resilient and healthy campus environment.