Research
at U of M
In the wake of the COVID-19 crisis, research has turned it's gaze to help fight this pandemic. Many efforts have been made at the University of Michigan to understand and model the pandemic and it's spread, specifically the Complex Systems department. These faculty members and their contributions to modeling and understanding the effects of COVID-19 will be featured on this page.
Dr. Marisa C. Eisenberg
Dr. Eisenberg is a faculty member in Epidemiology, Complex Systems, and Mathematics and has recently been featured on the Complex Systems website for her work in modeling the spread of COVID-19 in the state of Michigan and created a website called the Michigan COVID-19 Modeling Dashboard.
The front page has current forecasts for 5 different scenarios: 1-week forecast of cumulative cases in Michigan, 3-week forecast of cumulative cases in Michigan, 3-week forecast of cumulative deaths in Michigan, 3-week forecast of cumulative hospitalized patients, 3-week forecast of ICU occupancy, 3-week forecast of patients needing ventilators, and 3-week forecast of patients needing O2 support.
The graph to the left is the 3-week forecast of cumulative COVID-19 cases in Michigan as of April 19th. The black dots are actual data points, where the grey shaded area are uncertainty bounds of 95% for 1000 simulations. The latest data point is 31,068 cases, while the uncertainty bounds are 28,383 to 53,826. These models and predictions are being updated daily as the pandemic proceeds.
Another informative feature is Dr. Eisenberg's interactive graph that demonstrates social distancing and its impact on how the pandemic spreads. Above is a capture is an example of what the effect social distancing has on the spread of the COVID-19 virus. This graph allows users to see what "flattening the curve" looks like and why early social distancing before the peak is effective. Dr. Eisenberg outlines that starting social before the peak will give the best results of the the method. Additionally, keeping the social distancing after the full peak lowers the risk of a "rebound" or resurgence of cases.
Finally, on the About tab talks more in depth of the type of model Dr. Eisenberg used in this application. A more complex version of the SIR model (susceptible, infectious, recovered), which is explained in more depth on the Models fo COVID-19 page of this website. The expanded version of the SIR model can be seen in the diagram to the right. Additional factors that have been considered in this model are having the mild disease and not seeking care, having the mild disease and seeking care, having the severe disease, being hospitalized, and death. These factors take in account more of the realistic stages of response of the disease and also accounting what kind of need there will be for beds in hospitals and the like.
