
As the COVID-19 pandemic unfolds, epidemiologists have released several statistical models that have projected how the virus could affect countries across the globe in the coming weeks and beyond.
Dr. Linn Carothers, professor of mathematics at California Baptist University, has monitored the COVID-19 projections throughout the pandemic. He said that governments have mainly used models from the Medical Research Centre for Global Infectious Disease Analysis at Imperial College London and Oxford University as resources.
The study completed at Imperial College, led by epidemiologist Neil Ferguson, was originally released March 17. Carothers said this model projected outcomes based on behaviors and actions taken by governments and individuals to restrict the spread of the virus. This model projected up to 2.2 million deaths in the United States as a result of COVID-19 if the government did not take any measures to limit the spread of the virus. The model indicated that strict social distancing regulations would lower healthcare demands and overall death rates.
Carothers said these projections helped lead the government to enact stricter social distancing measures, and now the projections have lowered to 60,000 deaths instead of 2.2 million.
“The model has had a positive effect in the sense that it has caused people to act and to do whatever we can do to help suppress the effects of the model,” Carothers said.
In contrast, the Oxford University model suggests that much of the population could have been exposed to COVID-19 and that herd immunity could have occurred, meaning that much of the population could have been exposed to and become immune to the virus already.
Carothers said that these two models offer different projections because they use different parameters and infection rates and because they have different aims.
“The math in both of them is good, but you have to realize that these two major models that people have looked at are actually different in terms of what they are asking,” Carothers said.
The Imperial College model aims to study how individuals’ behavior can affect death rates and healthcare needs, while the Oxford University model focuses on immunity and how COVID-19 has affected different groups. However, Carothers said both models can offer helpful information during the pandemic.
Sam Robertus, freshman sports analytics major, said, “Models and projections can help both officials and the public know what to expect in this unpredictable time and can help give some much-needed security of at least having an idea of what the virus might do. It also helps government officials establish proper guidelines and regulations.”
Models also offer projections about the amount of medical equipment and personnel a given area might require at a certain time during the pandemic. Anika Heintz, senior nursing major, said that projections can indicate potential issues that hospitals may face with sustaining the number of COVID-19 cases.
“Some hospitals are unable to keep up with the demand of patients and the lack of Personal Protective Equipment,” Heintz said. “These projections of a steep curve could overwhelm our healthcare system to a point where more people may unnecessarily die because of low numbers of ventilators and hospital beds. If we flatten the curve, hospitals can adequately care for these patients and lower the number of cases and deaths.”
Although models can provide information about how COVID-19 might affect people in the future, Carothers said that models will not be completely reliable during the pandemic.
“Models are not reality,” Carothers said. “Models are simply projections based upon assumptions of what we think might be the outcome. The models people have been hearing from different places are projections that are not going to be completely accurate, but at least they will give us a view of what the future might look like based on certain assumptions and things we either do or do not do.”
Carothers said predictive models are not entirely reliable because many factors can change because of people’s behavior, the government’s response and the possibility of mutations of the virus. In addition, the current projections are based on equations that take into account categories such as people who are susceptible to the virus, exposed, infected and recovered. Carothers said these models are less accurate than more complex ones that require data about individual behavior and movements of people.
“At the moment, we are missing actual data on transmission and for us to get accurate answers for the factors we need in the model, we are going to need large-scale antibody testing,” Carothers said. “That would at least allow us to determine levels of herd immunity, how many people have come in contact with it and how many recovered, which we do not know.”
The University of Southern California announced a new antibody test for COVID-19 on April 15. Carothers said he hopes that as more data becomes available as the pandemic progresses, projections will become more accurate. However, Carothers said he believes the current models can still be used to make a difference and raise awareness about the virus and its impacts.
“I think the models are very helpful in that it was the models that moved the governments to action,” Carothers said. “They are helpful in helping individuals and households plan their activities and realize how important it is for them to listen and follow the current suggestions. Mathematical models are helping us understand how our personal behavior can affect our entire community and nation.”