Since the coronavirus hit the United States, authorities and citizens have measured the severity of the spread by tracking one particular measure: how many new cases are confirmed by testing each day. However, it was clear all along that this figure is an understatement because of the gaps in the tests.
Now, a research team at Columbia University has built a mathematical model that gives a much more complete – and frightening – picture of how much virus is circulating in our communities.
He estimates how many people are never counted because they never get tested. And it answers a second question that is arguably even more crucial – but which has so far not been reliably estimated: on any given day, what is the total number of people actively infectious? This includes those who may have been infected in the past few days, but are still spreading viruses and are able to spread disease.
The conclusion of the model: on a given day, the actual number of active cases – people newly infected or still infectious – is probably 10 times the official number of cases reported on that day.
The model has not yet been published or peer-reviewed, but lead researcher Jeffrey Shaman, an infectious disease specialist at Columbia University, shared the data exclusively with NPR. Here are more surprising tips.
Lost cases remain a major problem
To arrive at a final estimate, the researchers’ first step was to estimate, for each day of the outbreak so far, how many people have actually become infectious. Then they compared this to the number of people tested and counted as a confirmed case.
This discrepancy alone was huge: the Shaman estimates that during the entire pandemic, five times more people were infected than reported.
“The numbers are greatly amplified,” says Shaman. “When we look at confirmed cases, we are seeing only the tip of the iceberg.”
The test rate in the USA has improved over time. The Shaman model found that, at the beginning of the pandemic, only 1 in 10 cases were being reported. In early May, it had risen to 1 in 6. In September, it was 1 in 5.
The Shaman estimates that, on average in the last three months, the official count counts only 1 in 4 infections. In other words, says Shaman, to get a rough idea of the actual number of new cases per day, you must multiply the number reported daily by four.
It gets worse – after considering the current active infections
Even estimating the actual number of new infections daily does not provide a complete picture of how risky it can be to mix with people in your community right now.
The estimated Shaman numbers for how many people became infected each day only tell who a new case is. But people remain contagious for “three or four days on average,” says Shaman.
Therefore, to fully assess the threat level on any given day, you would also like to count the people whose infection started earlier and who are still transmitting the virus.
“There are a lot of people out there with this virus who never know they have it,” says Shaman. “Even people who are tested and confirmed, they were contagious before they even had symptoms.”
So this is the next step for the Shaman model: he estimates that the number of people actively spreading viruses on any given day is about 10 times the number of new cases reported daily.
How many people does this add up to? Well, on the worst day for new cases reported 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 were actually actively spreading the virus.
The peak was even worse in many jurisdictions. In Los Angeles County, says Shaman, in the dead of winter, 3% of the county’s population was contagious, or about 3,000 per 100,000.
Transmission has declined considerably since then in the United States. But it is still well above the summer highs. And Shaman estimates that as of last Saturday, 1.25 million people across the country were actively spreading viruses.
“This is a very, very high level,” says Shaman. “It still means that there are many people who are actively infected, who are transmitting the virus and who can expose people at risk.”
Why does that mean we can’t hurry to open
The findings give urgency to the rush to vaccinate Americans, says Shaman. And it suggests that Americans will need to maintain a high degree of physical distance and masking until many more people are vaccinated.
“If we stop now, given the amount of infection that exists, we are going to get many, many more people to contract the virus before they even have a chance to get the vaccine,” said Shaman.
Ashish Jha, a public health researcher and dean of the Brown University School of Public Health, says he considers these new models “really important”, although he cannot assess the methodology of the model since it is not public.
“What people really care about is not: ‘How many people in my city or state became an affair yesterday?’ “says Jha.” Yeah, ‘When I’m out and about, how many people around me are infectious? How many people around me are potentially spreading the virus?’ This is the first [work] I saw that you really try to get there. “
One third of the US population has already been infected.
Sustained periods of high transmission in the United States also mean that a large part of the United States’ population has now been infected beyond what the reported case counts would indicate. Across the country, Shaman estimates that about 120 million people have already been infected, just over a third of the US population.
The model also provides estimates for each state.
There is great variation: in North Dakota and New York, for example, Shaman estimates that about half the population has already been infected. “They may even be approaching collective immunity there,” he says.
But Shaman also warns that it is possible that the immunity acquired from the infection – especially in mild or asymptomatic cases – may decrease before enough people are vaccinated to contain the outbreaks. It is also unclear what degree of protection the previous infection will confer against some of the new variants that have recently been detected in the United Kingdom, South Africa and Brazil – and which many scientists assume will become increasingly common in the USA.
In addition, in many states, the proportion of people infected is much lower. And the United States in general – with an estimated one-third infected – is nowhere near the 70 to 85 percent level that scientists estimate must be immune before the pandemic starts to subside here.
Shaman’s conclusion: “I don’t think we should think psychologically of any kind of move to a post-pandemic phase and a real reopening until the summer.”
“The important thing,” he adds, “is not to be overly exuberant now and think that we are done with it.”
How does this model compare to previous estimates
The Shaman is not the first to try to estimate how many infections were lost in the tests. This part of his analysis – although not modeling the total number of active infections – echoes previous research.
Shaman found that, at the start of the pandemic test, it caught only one in 10 new real infections – this is in line with estimates by researchers at the Centers for Disease Control and Prevention in the United States. These studies include several extrapolated blood samples that looked for antibodies to the coronavirus – which is evidence of previous infection. They suggested that the number of actual infections was 10 times higher than reported.
In another study by researchers with the CDC, a similar model, but more rudimentary than the version used by Shaman’s team, found that actual infections were probably eight times greater than reported during the first seven months of the pandemic.
So, how did the Shaman team arrive with their estimates? They started with two known pieces of information: the first was the number of people who tested positive each day since the pandemic began. The second was a set of anonymous cell phone location data – provided by the SafeGraph company – that reported, each day, how many people were mixing up when moving out of their homes, including, says Shaman, “to places of interest like supermarkets and restaurants. “
The team then fed this data into a computer program that essentially tried to find the best possible answer to the variables whose value the team gave no you know – things like, how many cases were lost each day? And how long did people stay infected?
The program effectively ran several simulations to see, for each day of the pandemic, which combination of responses allowed to correctly predict how many reported cases were produced in the following days. In a nutshell, says Shaman, “he seeks the ideal solution that best fits observable data.”
Sydney Lupkin of NPR contributed to this report.
Copyright NPR 2021.