IBM Model for a Covid-19 Outbreak in College Communities

We simulated an outbreak of Covid-19 on a college campus using a discrete, individual-based model of disease transmission. Our model uses a detailed account of random and scheduled student contacts as well as a meticulous simulation of a Covid-19 infection to model disease transmission and progression accurately. The goal of our work was to create this web-based application modelling Covid-19 on college campuses such that college administrators or other interested parties could test different disease scenarios and interventions without knowledge of programming.aBelow are the model parameters that set up a college environment and simulate disease spread in that environment. There are three classes of parameters: Initialization Inputs, College Inputs, and Disease Inputs. Every parameter in every category has a default value and if necessary a short description explaining what the parameter means to aide in customizing your simulation.a Initialization InputsWith the same college community and under the same disease transmission and progression assumptions, the model gives the option to run simulations that differ in the number of students in the model, the number of weeks the model is run for, the fraction of the population that is immune, and the number of students who begin infected.aCollege InputsUsing a few key inputs you can create an existing college or you can generate a synthetic college. The inputs that control the college community and student interactions are mean class size, minimum relative preference level of classes, proportion of MWF to TTH classes, fraction of student body that lives in the dorms, number of students allowed to share a university residence, and the size and relative frequency of social mixing groups.aDisease InputsThe disease inputs are the inputs to the model that control how the disease progresses within a host and how it spreads to infect new hosts. These parameters are: fraction of the population that is in the symptomatic disease pathway, the maximum infectivity and duration of infection for members of both the symptomatic and asymptomatic pathways, the environmental risk factors of the class, dorm, and social mixing environments, and the pathogen spread length scale in classroom.aTroubleshootingWe are very sorry if you experience difficulities running the model. As a web app it faces time and memory limitations that normal Mathematica programs do not face. We appreciate your patience as we continue to work to make this model implementation both convenient and powerful.a If you receive an error code that says the memory limit has been exceeded, we recomming reducing mean class size and/or number of students. We are working to have more memory available. If it is taking a long time to run, it is possible that someone else is simultaneously trying to use the web app. We recommend trying it at another time and/or reducing the number of students.a

The number of students in the model is the number of students in the college. We do not recommend doing more than 3000 students.

The model will track how many students are becoming infected or recovering over the duration of the simulation. You could choose this duration to be, for instance, a semester.

The fraction of the population that is immune is the fraction of the population that cannot catch the disease. This can be used as a proxy for fraction of the population vaccinated. The key difference being that vaccines are less than 100% effective and this immunity is 100% effective. (Number between 0 and 1.)

The miminum relative preference level of classes is meant to capture how much variation in size there is between classes. If the most popular class has size 1, what size is the least popular class? Is it half the size or maybe a quarter of the size? Fill in that fraction. (Number between 0 and 1.)

The proportion of MWF classes to TTh classes is the number of MWF classes that occur for every TTh class. You can estimate this proportion or you can calculate it as the number of MWF classes that occur at the university dividied by the number of TTh classes that occur at the university.

The fraction of the student body that lives in the dorms is the fraction of the population that lives in university housing. (Number between 0 and 1.)

The number of students allowed to share a university residence captures in what size groups students live at the university. Fill in how many students share a dorm, a public utility, or are in contact in a residence setting according to university policy.

The pool sizes are the sizes of the groups in which students gather socially

The relative frequeceny of these groups is how often they relatively occur. If one pool size has a relative frequency twice that of another pool size then it occures twice as often.

The pool sizes are the sizes of the groups in which students gather socially

The relative frequeceny of these groups is how often they relatively occur. If one pool size has a relative frequency twice that of another pool size then it occures twice as often.

The fraction of the poputlation in the symptomatic disease pathway is the proportion of students that if infected will experience symptoms. (Number between 0 and 1.)

The maximum infectivity of a person in the asymptomatic pathway is how infectious they will be when they are at their most infectious. (Number between 0 and 1.)

The maximum infectivity of a person in the symptomatic is how infectious they will be when they are at their most infectious. This number will likely be larger than the maximum infectivity of a person in the asymptomatic pathway. (Number between 0 and 1.)

The duration in days of infection for a person in the asymptomatic pathway is how many days they remain infected once they catch the disease.

The duration in days of infection for a person in the symptomatic pathway is how many days they remain infected once they catch the disease. This number will likely be larger than the duration in days of infection for a person in the asymptomatic pathway.

The environental risk factor for class mixing is a multiplier which modifies how likely an infected person is to transmit the diease to a susceptible person in a classroom. This is a function of physical conditions and behavioral practices such as seat density and mask wearing. (Number between 0 and 1. 0 is no likelihood of tranmission. 1 is highest possible likihood of transmission. Suggested values for low, moderate, and high levels of risk are .1, .2, and .45 respectively)

The environental risk factor for dorm mixing is a multiplier which modifies how likely an infected person is to transmit the diease to a susceptible person in the dorms. This is a function of physical conditions and behavioral practices such as the specifics of the dorm layout. (Number between 0 and 1. 0 is no likelihood of tranmission. 1 is highest possible likihood of transmission. Suggested values for low, moderate, and high levels of risk are .1, .2, and .45 respectively)

The environental risk factor for social mixing is a multiplier which modifies how likely an infected person is to transmit the diease to a susceptible person in social interactions. This is a function of physical conditions and behavioral practices such as if they are gathering indoors or outdoors. (Number between 0 and 1. 0 is no likelihood of tranmission. 1 is highest possible likihood of trasnmission. Suggested values for low, moderate, and high levels of risk are .1, .2, and .45 respectively)

The pathogen spread length scale in classrooms is how quickly the likihood of infection decreases with distance in classrooms. A larger number signifies a faster decrease over distance and lower pathogen transmission. This parameter reflects real world quantities like air circulation in classrooms and is a measure of how easily pathogens carry across the room.