[ad_1]
At the end of February, there were fewer than 20 known cases of Covid-19 in the United States. But at a two-day conference for a pharmaceutical company in Boston, someone must have carried the virus. More than 90 of the 200 conference participants and their families were infected, and the virus continued to spread through Massachusetts and other states. Researchers estimate that tens of thousands of people were infected in and around Boston alone.
At a wedding in Jordan on March 13, the father of the bride infected at least 76 out of 360 guests with the novel sars-cov2 coronavirus. And yet. And on March 10, a person with cold symptoms infected 52 of the other 60 singers during a driving practice in the US state of Washington. Two of the choir singers died in covid-19.
Cases of super dispersion have been reported at parties, factories, gyms, cruise ships, churches, and other settings where many people gather indoors in a limited area for a long time.
Similar patterns exist for several infectious diseases, but for covid-19 it is clearer. A small proportion of those infected account for most of the spread of infection. In a study of two southern Indian states recently published in the journal Science, five percent of those infected caused 80 percent of the cases in the next stage. 70 percent of the sick did not infect anyone else.
Scientists don’t know why some people are much more contagious than others. Despite this, there are good conditions to prevent overproliferation opportunities from arising.
– Super scatter events depend on the interaction of several different factors. One is the biology of the infected individuals. But beyond that, most super-scatter events happen when many people gather for a long time in a small area. You could say that the environment itself is a super spreader, says Serina Chang of the Department of Computer Science at Stanford University.
She and her staff have created a computer model of how the infection spread in various places, including restaurants, hotels, gyms, churches, and grocery stores, in major cities across the United States during the spring.
– Ten percent of locations accounted for more than 80 percent of infections in these metropolitan areas between March and May, according to our results. It is quite remarkable. Therefore, the infection is not spread evenly in society, says Serina Chang.
Restaurants with table service had the largest environmental spread in the researchers’ model, according to results published this week in the journal Nature.
– Restaurants are four times more dangerous than the next category, which is the gym. Then come the cafes and hotels, says study leader Jure Leskovec, an associate professor of computer science at Stanford.
The researchers used mobile phone data that shows how people move between their residential areas and different public settings.
– We use mobile data to simulate the spread of covid-19. We captured the pattern of movement of nearly 100 million people in the ten largest metropolitan areas in the United States, says Jure Leskovec.
The new model may help decision-makers deal with the pandemic, the researchers hope.
– Eight months after the pandemic, there is still a debate about when we should open society again, which places should be opened and how we should open them. We really believe that a stronger empirical basis is needed to choose the right strategy, says Jure Leskovec.
– According to our results, we should not treat all different environments the same, nor open them in the same way, because the risk of spreading infections varies a lot, says Serina Chang.
Mobile data The researchers had access to it, it was anonymized, but they could divide it into groups of between 600 and 3,000 people according to the different residential areas on which people’s daily movements were based.
– We could model how people move hour by hour and predict where and when someone would be infected, says Serina Chang.
The researchers assumed that there was a small proportion of those infected in all groups at the beginning of the pandemic.
– Our simulation starts on March 20. Then we press play. Every hour thereafter, a few people from the different groups head to one of the settings. All movements in the model are based on real mobile data. Some people are susceptible to viruses, others are infected. How likely someone is to get infected in the model depends on the size of the room, how long you are there, how many other visitors are there at the same time, and how many of them are infected, says Serina Chang.
Even when the people in the model are at home, there is a certain risk of getting infected.
– To calibrate and verify that our model is reliable, we compare our predictions with the actual number of daily cases of covid-19, according to the New York Times report. We show that our model can predict the number of reported cases in the ten metropolitan areas, says Serina Chang.
During spring received several model calculations with grim predictions for the development of the pandemic in Sweden received much attention. With the results in hand, it turned out that the models had overestimated both the need for intensive care units and the large number of deaths. According to a study recently published in the Journal of the Royal Society Interface, many of these models end up in real trouble because they expect a uniform spread of infection and do not take into account the dynamics of super spread events that have been shown to be so characteristic of covid-19. The Swedish models were also criticized early on because they were too complex and contained a large number of unknown parameters.
In contrast, the model that Serina Chang and her co-workers have created has very few parameters.
– Mobility is the only parameter that varies over time. During this period, from March to May, people’s behaviors changed a lot, for example, with mouth protection, hand washing, social distancing and the like. The fact that our model is still so successful in recreating the actual infection curve may indicate that mobility, in particular, plays a crucial role in the spread of infection, he says.
All over the world the pandemic has hit low-income people and other vulnerable groups the hardest. The researchers’ model offers one possible explanation: vulnerable groups have not been able to pay or have not had the opportunity to stay at home.
– It is striking that our model makes correct predictions that groups with lower incomes and with a lower proportion of whites are infected with covid-19 at a significantly higher rate, only based on mobility. This is probably due to the fact that groups are overrepresented among people who have jobs that cannot be managed from home. It makes a difference whether you work or just shop at a grocery store, for example, says Serina Chang.
– The grocery stores that low-income people visit are generally more crowded. According to our model, a visit to the grocery store is twice as dangerous for a low-income person as it is for a high-income person, says Jurek Leskovec.
Serina chang
Read more: When few infect many, over-spread can affect the pandemic
Read more: Reduced travel after sharp crown tips
