Since the coronavirus reached the U.S., officials and citizens have measured the severity of the spread by following one measure in particular: how many new cases are confirmed each day by testing. However, it was consistently clear that this number was an understatement due to the shortcomings.
Now a research team at Columbia University has built a mathematical model that gives a much more complete – and narrow – picture of how many viruses are circulating in our communities.
It estimates how many people are never counted because they are never tested. And that answers a second question that is probably even more important – but which has not been reliably estimated so far: what is the total number of people who are actively contagious on a given day? This includes those who may have been infected in the previous days but still shed viruses and can spread diseases.
The model’s conclusion: On any given day, the actual number of active cases – people who have just been infected or are still contagious – is probably 10 times that day’s official number of reported cases.
The model has not yet been published or evaluated by peers, but lead researcher Jeffrey Shaman, a specialist in infectious diseases at Columbia University, shared the data exclusively with NPR. Here are many more of the scrumptious takeaways.
Missing cases remain a major problem
To come up with their estimate of the conclusion, the researchers’ first step was to estimate for each day of the outbreak so far how many people had become contagious. Then they compared it with the number tested and counted as a confirmed case.
Just this difference was huge: Shaman estimates that five times more people are infected than have been reported across the entire pandemic.
“The numbers are amplifying a lot,” Shaman says. “When we look at confirmed cases, we actually only see the tip of the iceberg.”
The test speed in the US has improved over time. According to Shaman’s model, only 1 in 10 cases are reported at the beginning of the pandemic. At the beginning of May it rose to 1 in 6. By September it was up to 1 in 5.
Shaman estimates that the official count has averaged only 1 in 4 infections over the past three months. In other words, says Shaman, you need to multiply the daily reported number by four to properly understand the actual number of new cases per day.
It’s getting worse – once you consider current active infections
Even estimating the actual number of daily new infections does not give a complete picture of how risky it is to mix with people in your community right now.
Shaman’s estimated figures for how many people became contagious each day just tell you who’s a new case. But people stay contagious for an average of three to four days, Shaman says.
In order to fully appreciate the threat level on a given day, you also want to count the people whose infection started earlier and who are still shedding the virus.
“There are a lot of people around with this virus who never know they have it,” says Shaman. “Even the people who were eventually wiped out and confirmed were contagious before they even had their symptoms.”
So this is the next step in Shaman’s model: He estimates that the number of people who actively shed viruses on a given day is about ten times as many as the new cases reported daily.
How much is it? Well, on the worst day for reported new cases so far – January 2 – 91 out of every 100,000 people in the US tested positive. But Shaman estimates that on that day, 998 per 100,000 people actively shed the virus.
In many jurisdictions, the peak was even worse. In Los Angeles County, says Shaman, 3% of the county’s population was contagious at the height of the winter storm, or about 3,000 per 100,000.
Shipping has declined considerably in the United States since then. But it is still far above the summer highs. And Shaman estimates that 1.25 million people nationwide have been actively shedding virus since last Saturday.
“It’s a very, very high level,” says Shaman. “It still means there are a lot of people who are actively infected, who pass it on and can put people at risk.”
Why it means we can not open fast
According to the findings, the rush to vaccinate Americans is urgent, Shaman says. And this suggests that Americans will have to maintain a high degree of physical distance and masking until many more people are vaccinated.
“If we’m disappointed now, given how much infection there is, we’re going to make it so that many, many more people are going to get the virus before they ever get the chance to get the vaccine,” says Shaman.
Ashish Jha, a public health researcher and dean of Brown University’s School of Public Health, says he considers these new models ‘very important’, although he cannot assess the methodology of the model as it is not public. .
“What people really care about is not, ‘How many people in my city or country became an issue yesterday?’ says Jha. “It’s, ‘How many people around me are contagious when I’m on the road? How many people around me are the virus possibly spreading?’ This is the first [work] I saw that it was really trying to get right. ‘
One-third of the U.S. population is already infected.
The sustained periods of high distribution in the US also mean that a fairly large portion of the U.S. population is infected by this time above what the numbers of reported cases would indicate. Nationwide, Shaman estimates that about 120 million people are now infected, just over a third of the U.S. population.
The model also provides estimates for each state.
There is quite a variation: in North Dakota and New York, for example, Shaman estimates about half of the population is now infected. “They can even approach herd immunity there,” he says.
But Shaman also warns that it is possible that the immunity gained from infection – especially due to mild or asymptomatic cases – may decrease before enough people are vaccinated to stop outbreaks. It is also not known what degree of protection pre-infection will provide against some of the new variants that have recently been detected in the UK, South Africa and Brazil – and which, according to many scientists, are increasingly found in the USA.
In many states, the proportion of infected people is also much lower. And the United States – with an estimated one-third infected – is nowhere near the level of 70 to 85% that scientists say must be immune before the pandemic can begin to subside here.
Shaman’s conclusion: “I do not think we should think psychologically of any form of transition to a post-pandemic phase and a real reopening to summer.”
“The most important thing,” he adds, “is not to get too extravagant now and think we’re done with this thing.”
How this model can be compared to previous estimates
Shaman is not the first one to try to estimate how many infections are missed. This part of his analysis – although not the modeling of total active infections – reflects earlier research.
Shaman found that early in the pandemic test, only one in ten actually caught new infections – this is in line with estimates from researchers at the U.S. Centers for Disease Control and Prevention. These studies include several that have been extrapolated from blood samples that were looking for antibodies to the coronavirus – evidence of previous infection. They suggested that the number of actual infections was ten times higher than reported.
In another study by CDC researchers, a similar but more rudimentary model than the version used by Shaman’s team found that actual infections were eight times higher than reported during the first seven months of the pandemic.
So, how did Shaman’s team come up with their estimates? They started with two pieces of well-known information: First, the number of people who tested positive every day since the onset of the pandemic. The second was a set of anonymous cell phone location data – provided by the company SafeGraph – that told them for each day how many people mingle by moving outside their homes, including, Shaman says, “to places of interest such as grocery stores and restaurants . ‘
The team then enters this data into a computer program that tries to find the best possible answer to the variables that the team has the value of. not know – things like how many cases are missed every day? And how long have people remained contagious?
The program effectively performed several simulations to see, for each day of the pandemic, what combination of responses enabled it to correctly predict how many reported cases were produced in the days that followed. In a nutshell, Shaman says, “it seeks the optimal solution that best matches the observable data.”
Sydney Lupkin, NPR, contributed to this report.
Copyright NPR 2021.