This page collects links to stories of surveys and universal tests for SARS-CoV-2.
(UPDATE) New York Antibody Test
Universe: New York State
Population: 19.44 million
Sample size: 3000
Infection rate: 13.9% (2.7 million)
Known cases: 263632
Infection fatality rate: 0.5%
Case fatality rate: 5.9%
These are only hospital and nursing home deaths. Deaths at home were not recorded. (This means CFR and IFR will be higher.)
“From March 8 to April 11, 2020, there were 15,100 more total deaths than average in New York state over the same period over the past five years, according to the CDC. According to The New York Times, 12,686 deaths from coronavirus were reported in NYS over those five weeks — a difference of more than 3,400 people.”
This number includes all kinds of deaths, not just COVID-19 deaths. The assumption is that most of them are COVID-19.
The key data: 21.2% of New York City residents tested positive for COVID19 antibodies. 16.7% for Long Island; 11.7% for Rockland and Westchester (the suburbs just to the north of the city); and 3.6% in the rest of the state
Chelsea, Massachusetts, 32% had Antibodies
Chelsea is a city of 40,000 people, just across the Mystic River from Boston, and has one of the most severe outbreaks in the state.
Wikipedia on Chelsea: Chelsea is a diverse, working-class community that contains a high level of industrial activity. It is one of only three Massachusetts cities in which the majority of the population identifies as Hispanic or Latino, alongside Lawrence and Holyoke.
Los Angeles County, CA, estimated 4.1% have antibodies
Date of Study: April 10 & 11 Location: Los Angeles County, CA Population: 10105518 (2018) Sample Description: Random, LA County, from LRW Group marketing database. Sample Size: 836 Test Description: rapid antibody serology, manufacturer not stated Infection Rate Estimate: 4.1% of LA County has antibodies Measured Infection: April 10, total 8430 (From LA Public Health Website, includes LB and Pas) Measured Deaths: April 10, total 241 Estimated Antigen Tests: April 10, total 40700
On April 10 and 11, USC ran rapid antibody test on a set of people forming a representative sample of the local population (from a marketing company). The press release didn’t note the number of participants. (I think these were the tests Los Angeles bought from South Korea.)
“Based on results of the first round of testing, the research team estimates that approximately 4.1% of the county’s adult population has antibody to the virus. Adjusting this estimate for statistical margin of error implies about 2.8% to 5.6% of the county’s adult population has antibody to the virus- which translates to approximately 221,000 to 442,000 adults in the county who have had the infection. That estimate is 28 to 55 times higher than the 7,994 confirmed cases of COVID-19 reported to the county by the time of the study in early April.”
So, we are very far from herd immunity. Wikipedia’s Herd Immunity page says we would need 24% to 79% of people to be infected before it stopped the spread of SARS-CoV-2. I have no idea how they came up with these numbers, but there’s such a wide range.
4% of people have the antibody that might make them immune to reinfection.
Santa Clara County, CA, ESTIMATED 2.5% to 4.2% have antibodies
From LiveScience: “That’s the takeaway from a small study of coronavirus antibodies in more than 3,000 people in Santa Clara County, California. The results suggested that between 2.5% and 4.2% of people in the county have contracted COVID-19, which is 50 to 85 times greater than the number of cases being reported at the time. Not everyone is convinced the true prevalence is that high, however, with some saying the antibody test the researchers used was not reliable.”
This sample was drawn from responses to Facebook ads, so there’s a self-selection bias favoring people who click on Facebook ads. Commenter mendel helpfully provided the ad:
Ay ay ay. I think this ad is a problem. If they had said “random disease study” and omitted the disease name, I think they might have been able to assert that bias was controllable.
There’s a lot of debate in the comments. I don’t have the understanding to evaluate those statements, but there appear to be some problems.
However, the estimate of spread was similar to the numbers in Los Angeles, which was detailed in the LA Times story. So, maybe it’s not that far off, if the LA study is solid.
rerutled’s comment on the Santa Clara study seems to cast shade on the LA study as well. The sample size seems to be too small to use a test with this much error.
If these surveys are not accurate, they still show that the spread of virus is not that vast, but many multiples larger than the recorded cases, and the sample size needs to be much larger, meaning we need a lot more testing capacity. In the big picture, every piece of data is useful, but you are getting only a limited view on the entire situation.
If these two antibody surveys are roughly correct, then the infection fatality rate drops to around 0.1% to 0.2%. Unfortunately, we’d still need to get to herd immunity, which just seems so far away. [This was corrected on April 28.]
Also, please note that the infection fatality rate of flu not 0.1%. That’s an estimated case fatality rate, based on deaths/estimated illnesses. Like COVID-19, many people get infected with influenza and are asymptomatic. So the infection fatality rate of flu is < 0.1%.
Boston’s Pine Street Inn Homeless Shelter, 37% infected, No Symptoms
A “universal test” is when every person of a defined population is tested. Universal testing was conducted at a homeless shelter.
From Boston25: “Of the 397 people tested, 146 people tested positive. Not a single one had any symptoms.”
That means 37% were infected, which is every third person.
This was not an antibody test, so it’s possible that some people had antibodies, and they didn’t get infected, because they were infected in the past. So, it’s not telling the spread of the disease in the past.
I suppose you could take 4% of 397, and assume that 16 additional people had already come down with COVID-19. It’s a stretch, though. The homeless population differs from the general population with regard to living conditions, obviously, and cannot “shelter in place”. So let’s not assume that 16 had it.
What I find surprising is that they didn’t have symptoms. So, this may mean that people living on the streets have an edge against COVID-19. On the other hand, being in a shelter, and enclosed space, may have contributed to the spread of the virus.
Universal Screening for SARS-CoV-2 for Women Admitted for Delivery, 15.3% had SARS-CoV-2
This is a universal test where the universe is pregnant women admitted for delivery to an NYC hospital.
“Between March 22 and April 4, 2020, a total of 215 pregnant women delivered infants at the New York–Presbyterian Allen Hospital and Columbia University Irving Medical Center . All the women were screened on admission for symptoms of Covid-19. Four women (1.9%) had fever or other symptoms of Covid-19 on admission, and all 4 women tested positive for SARS-CoV-2 (Figure 1). Of the 211 women without symptoms, all were afebrile on admission. Nasopharyngeal swabs were obtained from 210 of the 211 women (99.5%) who did not have symptoms of Covid-19; of these women, 29 (13.7%) were positive for SARS-CoV-2. Thus, 29 of the 33 patients who were positive for SARS-CoV-2 at admission (87.9%) had no symptoms of Covid-19 at presentation.”
(29 + 4) / 215 = 15.3% of this population had SARS-CoV-2.
Again, the vast majority, 87.9%, didn’t have symptoms.
15.3% is way higher than the 4% or so estimated in LA and Santa Clara counties.
So if you want to just let the virus run free, you might experience something like NYC, and still end up without herd immunity.
Union Rescue Mission, Los Angeles, 23% had SARS-CoV-2, Some with Symptoms
200 tests, 187 individuals, 43 positive, 16 symptomatic.
100% of people, 23% positive, 9% symptomatic.
That means 27 (14%) are asymptomatic.
Of the positives, 37% were symptomatic, and 63% were asymptomatic. This stat is relevant to understanding how the virus affects this specific population. Of course, with time, more may show symptoms.
The oldest universally studied population. These people were on a ship, and infected each other. Some people still have active cases, but most have recovered. This site has some stats and time series plots.
The Diamond Princess had 3,711 passengers and crew. 712 got infected (around 20%). So far, 13 (2%) have died. Around half were asymptomatic at time of testing.
I’ll need to spend some time parsing this article, but you can divide up the prisons into those that have tested everyone, and those who test only people with symptoms.
- The way to calculate the spread of the virus is to do a random survey of a population, using an antibody test.
- You can’t really go from the antigen test and extrapolate to the population, because the population being tested is self-selected, and then screened to ration testing.
- The current criteria for antigen tests is typically as follows:
- The person has symptoms, from a list of symptoms.
- OR the person has been in close contact with a person who has tested positive.
- In the past, being in a high-risk category: older person, or immunocompromized person.
- In some situations, even having the symptoms is not enough, because the provider requires the person to have relatively severe symptoms! (This is happening with LA County public health hospitals.)
- This restrictive criteria for testing causes the proportion of positive test results to be higher than it would be in the general population.
- It’s possible that restrictive criteria for testing causes the number of deaths from COVID-19 to be undercounted.
- The current criteria for antigen tests is typically as follows:
- You can’t really go from a universal test of a specific population, and extrapolate to the entire population.
- It’s not a random sample that finds people in the general population.
- The above tests had various selection criteria:
- Pregnant women ready to deliver a baby. This is all women, and generally younger than average.
- Prisoners. Mostly men, confined in close contact with others.
- Homeless shelter. People confined in close contact with others. Lots of outdoor exposure.