[ad_1]
SCIENCE – This is the billion dollar question, if not a lot. Where are we really at greatest risk of being infected with the coronavirus? If we had the answer, we would know how to properly deconfine to avoid a third wave of Covid-19 while allowing life (and the economy) to resume.
It is therefore logically a very important but very complicated research axis. Several specific case studies made it possible to formulate hypotheses. By cross-checking them, we are able to draw a robotic portrait of the worst places that favor infection: closed, poorly ventilated, overcrowded. But it remains limited. And if precise epidemiological studies are underway, with a follow-up of hundreds of people, they are not perfect and, above all, they take time.
In works published this Tuesday, November 10 in Nature, American researchers have tried a new approach to answer this question. They analyzed the movement of populations in 10 major American cities using (anonymized) hourly geolocation data of 98 million people. They then used it to analyze the evolution of the epidemic from March to May, during the first wave that hit the country.
A mathematical model both very simple and very precise that confirms many things and teaches us new ones, estimates epidemiologist Marc Lipsitch in a related article published in Nature. Restaurants, gyms, cafes, bars, hotels and places of worship are places of welcome for the public where the risk of contagion is greatest, according to the study. And limiting the indicator would be the most effective measure to stem an epidemic recovery.
Obviously this job isn’t perfect. Like all models, this too is theoretical and possibly partial. Furthermore, the children could not be incorporated, which sheds no light on the school debate. But the findings could at least help health authorities better manage deconfinement to avoid future waves.
Simple and complex
To arrive at this result, the researchers then analyzed the movement of millions of Americans in 10 major cities which they divided into neighborhoods. They then observed how people move hour by hour through hundreds of thousands of “places of interest”. This represents 5.4 billion “points” per hour. Of course, travel has been drastically reduced with the measures taken by the authorities as the coronavirus advances in the United States.
Next, the researchers used a mathematical model to predict the course of the outbreak. The problem with the models is that you have to make assumptions about the spread rate, the people met, the effect of this or that measure, etc. The authors of the study therefore chose a very simple model, based on what we know best: the basic reproduction rate (R0).
They then adapted it geographically. Schematically, they made sure to match the evolution of the trip and that of confirmed coronavirus cases in each city. Each time, this model adapted to citizens’ movements has predicted the evolution of the epidemic from March to May better than the more classic models. Once their theory was validated, the authors examined in detail these points of interest that were most involved in the spread of the Covid-19 epidemic.
Places of super contamination laid bare
In Chicago, for example, 10% of locations are responsible for 85% of the contamination. This confirms what we already know: many infected will not contaminate anyone or almost anyone, but some will create large clusters: this is what is called a supercontamination episode.
Restaurants, even more classic than fast food, are the worst places, four times riskier than gyms or even cafes. Then come the hotels and places of worship. Further on, there are doctors’ offices, grocery stores, and many other less risky stores.
Because? Here it is not even a question of mask or barrier measure. Because we are talking about a period in which these measures have been little or not implemented. The authors believe that it is above all the rate and duration of employment that are taken into consideration. In short: the longer you stay, the more risks there are. The more crowded the place, the greater the risks. The combination of the two is obviously the worst.
Furthermore, the authors also realized that, with the same type of place, there was greater contamination in the geographical areas where the average income was lower. We know that Covid-19 hits the poorest populations the hardest, but we don’t know exactly why.
Thanks to their data, the authors proposed two avenues. The first is that travel in these areas decreased less during containment in the United States (because statistically there are more frontline workers). The second is that the same type of place is more crowded in a low-income neighborhood than in a wealthy neighborhood. And people spend more time there on average.
“It’s twice as dangerous to go to a supermarket in a low-income neighborhood as it is in a high-income neighborhood, because there is 60% more employment and people stay there 70% longer,” Jure says. Leskovec, co-author of the study and a researcher at Stanford University.
Ways for a more efficient deconfinement
In addition to shedding light on the past, this model can above all give us ideas for future deconfinements. “If a minority of places of interest produce the majority of infections, reopening strategies must specifically target those high-risk places,” the authors write.
There is also the question of how to deconfine while maintaining certain limits. On this point the researchers carried out a test by arbitrarily modifying the parameters of their mathematical instrument. By limiting the maximum level to 20% in these high-risk areas, infections are reduced by 80%, while the overall frequency drops by only 42%. Clearly, avoid rush hours (the most dangerous) by leaving the shops open.
This of course is just a model. Now it is necessary to test it with more recent data (with those of this summer when infections have started to rise again in particular in the United States), or even on other countries if it is possible to have this level of detail in terms of position.
But, according to the authors, these first results may “guide decision makers” in the search for a decontamination that does not lead to a third or fourth wave. Jure Leskovec also explains that the next step is to develop tools so that health authorities can “test different reopening strategies”.
Serina Chang, co-author of the study and researcher at Stanford University, also specifies that it would be possible to improve the model to analyze the impact of universities. On the other hand, it is impossible to analyze other schools for reasons of unavailability and data protection. Researchers would also like to be able to verify the impact of the workplace in the spread of the epidemic, but that would require being able to distinguish between a simple visit to a company and a job on site.
See also on The HuffPost: how to control an epidemic, instructions for use
Source link
