Revisions to the Curve
Since we published ‘The Curve is Already Flat’ a month ago, new studies have suggested that the SARS-CoV-2 novel virus was spreading outside China long before a test kit was developed to detect it in Jan 2020. For example, Nevada public health officials confirmed that they're convinced some people had COVID in December 2019, and a new French study confirmed the novel virus was in community spread in France the same month.
Now that serological antibody tests have confirmed widespread rates of past infection among populations in Santa Clara CA, Boston MA, Oregon, pregnant women in New York City, and prisoners in Ohio, the hypothesis that COVID arrived in the USA much earlier than official reports claim is getting stronger.
Given that transmission takes place only in close quarters, it doesn’t make sense that the virus arrived in the US in mid-January and then became epidemic in urban US centers just six weeks later. For example, one Australian study found that classrooms alone are not favorable environments for transmission, implying that even closer contact is necessary -- the kind of close contact that can only be achieved between two people who live together in a house, apartment, dorms, prison cell, nursing home, or share densely packed train cars, buses, South Korean call centers, and long-distance international flights.
Although SARS-CoV-2 is contagious in crowded settings, transmission outside these settings is slow. Thus, there are two or more rates of transmission, depending on the contact network characteristics -- a low rate that exists outside crowded communities, and a much greater rate inside them.
As a consequence, we should expect to see slow growth of COVID when it is spreading at the low rate between close-contact communities, and a sudden spike in cases as soon as it penetrates a new close-contact community. That kind of spread doesn't result in a smooth, bell-shaped curve like the kind you see in a newspaper graphic. Instead, it will bump along at a slow burn, punctuated by spikes and troughs.
The problem is that the official story — that COVID arrived in the US with an infected Chinese traveler sometime in mid-January 2020 — relies on a smooth, bell-shaped rate of increase predicated on a single attack/transmission model that is now increasingly preposterous. For example, by 23 Jan 2020, when China suspended travel in and out of Wuhan, the virus was already epidemic there.
How could a viral epidemic somehow miss every one of the hundreds of thousands of Chinese residents entering the United States between 17 Nov, 2019, when the novel virus was first identified, and the travel suspension two months later that was intended to prevent exactly that route of transmission?
It is only with a two- or three- month head start that infection rates could grow from 0% to 25% of New York City.
This improved understanding of how SARS-CoV-2 spreads through different communities has important implications for policy options intended to slow the spread of a novel virus that already spreads slowly outside certain settings. For example, critics of Sweden’s approach (including some of our closest and most trusted friends) lambaste the liberal approach as reckless and unethical, but ignore the fact that Sinagpore’s indefinite lockdown isn’t workable, either. As new case data from Singapore demonstrates, lockdown only slows the rate of infection. Without immunity in their general population, Singapore will remain vulnerable to importing new outbreaks of coronavirus from infected, asymptomatic international travelers.
Because that’s how the coronavirus spreads.
"The virus resides in human beings, is carried by human beings, and is transmitted between human beings,"
- Woods, Seager, Alderson in When Can We Move Forward? (2020)
How Did COVID Get Here?
A phylogenic studies of the SARS-Cov-2 genome suggests that the outbreak began in China at some time between Sep 2019 and Dec 2019 -- but where would the early cases show up in US health data? If coronavirus did travel from China to the United States in Nov 2019 without causing a rapid escalation in acute cases, it must have left its signature somewhere in the travel and health data collected and reported by the United States.
For example, according to the US Department of Commerce, over 2.8 million Chinese residents entered the US in 2019. That’s more than 230,000 per month, and it doesn't count residents of other countries who visited China and then traveled to the US.
That's a lot of traffic between China and the US.
About two-thirds of Chinese residents who enter the US travel on “pleasure" visas, about one-fifth on students visas, and the remainder (an eighth) come for business.
Chinese Resident Rates of Entry to the US, by Type of Visa: - Pleasure: 1,893,480 (66.9%) - Student: 576,558 (20.4%) - Business: 359,932 (12.7%)
To test the hypothesis that many of these Chinese travelers were carrying coronavirus to the US, we must examine the health data available from that period.
Because there was no test for the novel virus until Chinese doctors released the RNA sequencing data 12 Jan 2020, there was no way to confirm the state of infection of any of these travelers by direct testing. And even after the RNA data did become available, Center for Disease Control (CDC) policies did not make testing available in the US for several more weeks, except under extraordinary circumstances. Thus, thousands of asymptomatic travelers might have been on planes bound from China, spreading the infection among their fellow travelers, or infecting those with whom they came in close contact after their arrival.
Assuming that some of these Chinese travelers become symptomatic shortly after their arrival in the US, they might seek treatment within the US healthcare system. Maybe the busy business travelers, who typically stay for shorter lengths of time, would decide to wait on seeing a doctor when they return home — especially when their symptoms are mild. And the pleasure travelers might be here for longer, but without access to or understanding of the US healthcare system.
The group of Chinese travelers most likely to report flu-like symptoms to US doctors are the student travelers. Because they’re in the US for longer and have access to health insurance from their US universities, they might be the first among Chinese travelers to seek treatment.
To find the students in the CDC flu data, we have to examine the way that ILI data is broken down into four age categories: 0–4 years, 5–24 years, 25–64 years, and 65+ years. International students might belong primarily to the 5–24 age group, but the prevalence of Chinese graduate students at US universities suggests that there might also be Chinese students in the 25–64 bracket. Therefore, the earliest detection of COVID in the CDC influenza-like-illness (ILI) data would have to be found in the 5–24 and/or the 25–64 year old age groups.
Which is exactly what the data indicates.
Figure 2 below shows the age breakdown on ILI in the United States from Fall 2019 — Spring 2020. Notice the 5–24 age group climbs first, dips at Thanksgiving, then shoots ahead by 10,000 visits a week more than the 25–64 year olds.
It shouldn’t be surprising to see a large number of ILI visits among the middle-aged adult group, given that they make up about half the total US population. However, it’s surprising to see that older children and young adults would be reporting flu-like symptoms more often than others, given that they’re only a quarter of the US population, and typically the demographic least likely to be impacted by flu.
The yellow line in Figure 2 above is the 65+ age group — the demographic most vulnerable to COVID. If this mysterious flu-like illness in Nov/Dec 2019 were COVID, then it should show up in the age group most vulnerable to it first, right?
Wrong, because the elderly live in the communities that are the last to be exposed, even if they are the first to die.
What we’re trying to find is not the cases that were spread inside the United States, but those that were imported to the United States. For example, over 10,000 Chinese people likely attended the Consumer Electronics Show (CES) in Las Vegas, Nevada, 7–10 Jan 2020. The Chinese auto industry (which has an enormous presence in Wuhan) was a major exhibitor. With COVID already epidemic in Wuhan by Dec 2019, it stands to reason that travelers to CES brought COVID with them.
Where would that show up in the ILI visit data?
It likely wouldn’t, until US residents infected at the show returned home, became symptomatic and reported to their doctor a week (or two) later. Which is (again) exactly what the data shows in the 5–24 and 25–64 age groups most likely to be participating at CES.
New data from autopsies in California support this thesis, revealing that the earliest known COVID death in the US was 6 Feb 2020 in Santa Clara, CA, one month after the CES. Since then, the Governor has ordered additional autopsies to determine whether an even earlier seed date might be supported by medical evidence.
Because fatalities are a lagging indicator of infections that were contracted in Nov 2019, then why wouldn’t we see excess deaths in the vulnerable 65+ age group by Dec 2019? That's what happened in the United Kingdom, where the only age group to experience excess mortality in the last two weeks of Dec 2019 and first two weeks of Jan 2020 were the 65+ year olds.
There could be several explanations for why the US didn't record a spike in 65+ fatalities until months after COVID began to spread through the University community.
The best of them may be this: COVID is idiosyncratic in that it spares the young. Because the first age group to report flu-like symptoms in the US were the young adults, and they bear little risk of fatality, it stands to reason that COVID wouldn’t show up in the mortality statistics until months later, after infected University students return home and infect their grandparents.
While none of this data is definitive, and it isn’t morally wrong for travelers to carry a disease they have no knowledge of, the thesis that COVID could come to the US in Nov 2019 without appearing in the fatality statistics is a compelling explanation for the widespread rate of infection in cities that house large populations of university students.
There’s just one piece of the puzzle missing: if COVID came to the US in young, otherwise healthy university students that are not at mortal risk, then shouldn’t we be reading anecdotes about real university students returning from China to a campus in the United States, infected with SARS-Cov-2?
That’s exactly what happened in my hometown, at the University where I teach.
Alison Nitzky contributed to this article.