One important finding: reopening the southern and western states led to a substantial drop in the “quarantine strength” .
Some of the studies listed in the study report have already been published, but haven’t yet been peer reviewed by experts in the area.
In the United States, the number of CoV-19 infections is growing rapidly, causing states to enforce quarantine measures and limit movement of the virus.
A model built by MIT researchers shows how effective a state’s quarantine measures are when based on the number of people infected.
The researchers identified their model in a study published in Cell, demonstrating the effect of the quarantine measures on the spread of the virus in regions around the world.
In their next report, recently published in the medical journal Preprint, they analyzed data from the U.S. last spring and summer.
The researchers found that the earlier increase in hospital acquired infections was closely linked to a decline in “quarantine strength” – a measure of the hospital’s ability to prevent people who are infected from infecting others.
The study investigates last spring and early summer, when there was a surge in infections in the southern and western U.S. as states reopened and tighter quarantine laws were relaxed.
The model was used to measure the severity of the quarantines in these states, which only reopened in the spring after an initial lockdown.
Had these states not reopened in early May, or if they had actually implemented interventions such as mask-wearing and social distancing, more than half of the infections may have been prevented.
Even after tightening quarantine measures, the report reports more than 100,000 illnesses may have been prevented in both Texas and Florida.
“When you look at these numbers, simple measures at the individual level can lead to a huge reduction in infection numbers and have a massive impact on the global statistics of this pandemic,” reports the research team.
As the country deals with an ongoing spring outbreak and states tightening restrictions, the team hopes the model will help policymakers assess the degree of quarantine steps.
“What I think we’ve learned quantitatively is that jumping around from hyper-quarantine to no quarantine and back to hyper-quarantine definitely doesn’t work,” said co-author Christopher Rackauckas, a professor of applied mathematics at MIT. “Instead, consistent application of guidelines would have been a much more effective tool.”
The MIT professors who co-authored the paper are also graduate students Emma Wang and George Barbastathis.
Nurturing our talents.
The researchers updated the SIR model, an epidemiological model used to determine how a disease will spread based on how many people catch the disease and get better.
Dandekar and his colleagues improved on an established SIR model by using a neural network that they learned to process data from real Covid-19 experiments.
The machine learning-enhanced model can estimate the number of infected individuals that do not transfer the virus to others and uses that knowledge to predict the spread of the disease (presumably because the infected individuals are following some sort of quarantine measures).
This value, referred to by researchers as “quarantine strength,” reflects how effective an area is at isolating infected individuals.
The model will process a region’s quarantine intensity over time to see a region’s evolution.
The model was created in February and has been used to forecast CCF in the early 2000s. The simulation was found to be indicative of the ground conditions in several European, South American and Asian countries that were initial hit hard by the virus.
“When we compare these countries to see when quarantines were implemented and compare that to the results for the trained quarantine strength signal, we see a significant difference.”