What kind of models are you using?
Depending on application, we are using several different types of models to understand and forecast COVID-19. For our statewide model, we are using a compartmental (SEIR) model that is specific for the 50 largest cities—by population size—in Idaho. Within each of the 50 cities, we break down the population by five different age categories in order to provide age specific estimates of morbidity and mortality. We are also using network based SEIR models to help model college campuses and agent based models to look at detailed intervention strategies within smaller populations.
What exactly is an SEIR model?
SEIR stands for a type of model that is appropriate for many respiratory illnesses. In these illnesses, the pathology of the disease is such that people start out as (S)usceptible, then become (E)xposed, then (I)nfectious, and finally (R)ecover. Being “exposed” is when an individual is “infected but not yet infectious”; this is often called a latent phase of the disease, where you are a carrier but not a risk to others at that time. Infectious individuals are those that actually spread the disease. Sometimes, if immunity is only short-term, we add another transition where recovered individuals become susceptible again. If this transition is included, we then call it a SEIRS model.
What exactly is an agent-based model?
In agent-based models, every individual is its own agent. Every agent has a specific behavior that they exhibit during a day or week. The agents also have their own set of traits such as age, sex, etc. These models take probabilities of interaction, infection, etc to create an explicit spatial and temporal pattern to life within, for example, a city. Then a disease is introduced into the city and allowed to spread. The very fine detail of an agent-based model allows us to assess specific, targeted interventions strategies that might be hard (or impossible) to model using an SEIR model. An example of such interventions might be something like closing specific stores (restaurants, bars) within a city.
Where do you get your data?
Data are an important part of calibrating our model to provide as accurate of a forecast as possible. We use data from multiple sources. In order to understand social distancing within Idaho, we rely on Google mobility data which are updated weekly. For travel between cities, we are using either geographic distance or travel patterns established by the Idaho Transportation Department. Our most reliable disease data are provided the Idaho Department of Health and Welfare and are not publicly available; however, we are also using more rapidly collected data streams within Idaho that are available on the state’s coronavirus tableau.
How up to date are the models?
Along with the information provided on the web site, we provide a date when a model was last updated. We try to continually update our models with the best possible data, but sometimes this may mean that the presented model is not as up-to-date as desired and results may need to be recalculated.
What do models really tell me?
Models are just that: models. Ideally, models mimic what is happening in the real world. However, there are many assumptions in models (in order to make them simpler) which can lead to inaccuracy. Typically, we like to think of forecasts from models as potential outcomes and not foregone conclusions. Given a set of assumptions, they are our best guess at what is happening and how that will continue in the future. Forecasts from disease models are no different than those for the weather; they are hopefully fairly accurate for the near future but are probably less accurate the further into the future you go.
What makes your model different from others?
In general, our models are not very different from many of the models out there. For example, many members of our team participate in the University of Texas COVID Modeling Consortium; at a basic level, our models are very similar to those that the UT consortium are using. Our models are unique in two ways. First and foremost, our models are tailored to the geography and demography of Idaho. This tailoring provides us with the best possible forecasts for Idaho, particularly when we wish to break Idaho down in different regions (e.g., public health districts). Second, we are focused on understanding and modeling the difficulties created by having an elderly, rural population. Such difficulties include things like access to health care in remote regions and increased vulnerability. These unique features are important to modeling COVID-19 in Idaho as accurately as possible.
Why model COVID-19 at all?
Models provide two critical capabilities. First, modeling COVID-19 allows us to predict what may happen with regard to the pandemic in the near future. These predictions allow the governor, public health officials, and private citizens to decide on the best course of action in the coming weeks. Second, models allow us to assess the potential impact of interventions (e.g., Idaho’s reopening phases) on the spread of coronavirus. This ability facilitates the decision making process; for example, models help decide when mask ordinances, social distancing, closures, etc. should be either put into place or lifted in order to provide the best possible outcome for the state.
General COVID-19 FAQs
What is COVID-19?
COVID-19 is a disease caused by a newly-discovered coronavirus. The virus, spread by close contact with an infected person, causes mild to moderate respiratory symptoms in most people. The most common symptoms are fever, cough, and tiredness. However, it can cause severe illness in some, especially older adults and people who have some underlying medical conditions. There is currently not a specific treatment or vaccine to treat COVID-19.
How can I protect myself from COVID-19?
The COVID-19 virus spreads through respiratory droplets spread by an infected person when they talk, sneeze, or cough. The Center for Disease Control (CDC) recommends that people: wash their hands often with soap and water for at least 20 seconds (or use hand sanitzer with at least 60% alcohol if soap and water aren’t available), avoid touching their face, avoid close contact with others by staying at least 6 feet apart from others, cover your face and mouth with a cloth mask when around others, cover coughs and sneezes, clean and disinfect commonly touched surfaces, and monitor your health. See more prevention tips from the CDC here.
Where can I get reliable information about COVID-19?
For general information about COVID-19, you can visit the Center for Disease Control (CDC) of the World Health Organization (WHO) online. For local information, check the website of your state, county, or health department. For Idaho, visit https://coronavirus.idaho.gov/.
Where can I find Idaho’s most up-to-date COVID-19 case data?
You can visit the Idaho Department of Health and Welfare’s COVID-19 Dashboard. It’s updated every day at 5 pm MT.
Where can I find information about the University of Idaho’s response to COVID-19?
Visit https://www.uidaho.edu/vandal-health-clinic/coronavirus to see what the University of Idaho is doing, including recent communications about current guidelines on campus