The Miami-Dade Test Results--And "My" Narrative

I'm taking the unusual step of pasting in a large portion of an article from the Miami Herald that provides the results of a serological (antibody) survey conducted in Miami-Dade County. I'm prefacing those excerpts with an attempt at a more systematic expression of my own views, or "narrative", as I've expressed them over the last few months.

My view has always been that, while the SARS-CoV-2 virus is almost certainly more dangerous than most seasonal flu (the deadliness of which varies with the season from year to year) once the disease is established in a person, whatever degree of deadliness it possesses will be mitigated by its lesser degree of virulence. Meaning, it is not transmitted as readily as are most strains of the flu. That view is not a result of my own insights, but rather expresses the early findings of scientists who have studied this virus. It also expresses my own seat of the pants view that the "human geography" of the US is different enough to provide a degree of protection when combined with the somewhat later appearance of the virus in the US.

And that means, in turn, two things--one good, and the other somewhat, well, mitigated:

1. We should not expect, contrary to what the "modelers" have presumed, that the virus to spread across the land like wildfire, and

2. On the other hand, contrary to what the "deniers" have maintained, the prevalence of the virus is very unlikely to reach rates anywhere near what is needed to establish a meaningful degree of 'herd' immunity.

I have also in the past featured articles that stress that the "dose" of the virus makes a difference. In other words, the vast majority of people are not going to develop the Covid19 disease just because they caught a whiff of the virus. The virus may be present, but the person will not develop the disease.

Old people in nursing homes and others who also have compromised immune systems, will be the most vulnerable in such a situation. Especially in nursing homes the inmates are being held in an artificially concentrated environment. Not only are they more inherently vulnerable, but they are more likely to receive a more direct and concentrated 'dose'. While you might think that cruise ships approximate that same situation, the different results on cruise ships point to significant differences. The population on cruise ships has been pre-screened for health and has usually been vaccinated for many diseases that might otherwise weaken their immune systems. They also have access to ample amounts of fresh air and sunlight.

In my view, the recent studies on the prevalence of the virus confirm my view. However, most of those studies have been pursuing what I regard as a wrongheaded approach. That approach is to maximise the instances of viral infection, no matter how minor, in order to minimize the "death rate." This is seriously misleading, because the tests that detect the presence of antibodies tell us little or nothing about the degree of immunity achieved (if any) or about how long any immunity will last--in the case of the common cold, another coronavirus, that immunity lasts for only about 2 months. Please note this, because it will be discussed in the the article below.

This misguided focus on "rates"--misguided, also, because bound to be an incomplete picture at this point--is not helpful in establishing policies that will both safeguard public health as well as enable the maximum of social and economic activity. Obviously I'm not enough of a specialist to present a coherent policy approach, but I do believe that the more common sense, not so statistical, view I've outlined offers a better starting point for discussion that could lead in a constructive direction, policy-wise.

Focusing on rates obscures the fact--which I have tried to draw attention to--that relatively simple precautionary measures, some of which were widely derided (inexpensive masks, gloves, common sense distancing, senior hour shopping, etc.), could be translated into measures applicable to high traffic, high density areas, such as public transportation and grocery stores. Many of those measures may still be highly advisable. The focus on rates and stats divorced from common sense understandings of the real suffering involved simply detracts from constructive discussion.

A final caution. I'm convinced that the evidence that this virus was engineered in a lab is strong. Given that dynamic--that the original bat virus was almost certainly modified to make it transmissible by humans and also to cause it to hide from the human immune system (which it does)--and the fact that viruses mutate rapidly, we do need to exercise caution. We simply don't know what direction this virus could take.

If I'm wrong, I'm wrong, but I'm not hiding my position--this post can be taken as a base line for my views. So ...

I think you'll see why I like this article. The authors of the study are open about their methods. They're open about the limits of their study--in fact, they draw attention to them. They don't try to obfuscate their results with statistical noise, and they didn't signal their desired results beforehand.



About 6 percent of Miami-Dade’s population — about 165,000 residents — have antibodies indicating a past infection by the novel coronavirus, dwarfing the state health department’s tally of about 10,600 cases, according to preliminary study results announced by University of Miami researchers Friday.

Please note: The percent needed for 'herd' immunity is usually calculated at 60%. There is no indication that the disease is spreading rapidly enough to reach that point any time in the foreseeable future. Also, health department stats vary widely from locality to locality. The fact that a low degree of prevalence--due, in my 'narrative', to a lower degree of virulence--"dwarfs" official statistics is really just 'statistical noise', confusing the discussion.

...

Friday’s results, based on two weeks of countywide antibody testing and about 1,400 participants, found that about half of the people who tested positive for antibodies reported no symptoms in the 14-17 days before being tested. If the trend holds, the findings could have major implications for understanding not only the number of people infected, but also how many have symptoms and, in turn, how the virus spreads.

To me, that's key: "how the virus spreads." The focus on this aspects is, to me, an indication of a seriously scientific approach, as opposed to a more ideologically oriented approach.

...

Gimenez said the UM survey of Miami-Dade reflects a national trend showing African Americans are disproportionately impacted by the novel coronavirus. “African Americans are more than twice as likely to be infected with COVID-19 than other racial groups,” he said.

The possibility of genetic factors being involved is, of course, a PC enforced third rail. We as a society need to get over that.

The survey findings also indicate that Miami-Dade is far from the 60% or higher infection rate needed to reach herd immunity, when enough people are infected to stop the spread of the virus by slowing transmission.

But this begs the question of just how transmissible this virus actually is. The widely expressed guideline--6 feet/10 minutes--suggests that it is not as transmissible as the flu or common cold. Many more anecdotal reports strongly support that conclusion. This factor needs to be part of the mix in considering what precautionary measures fit the case. My advice: Don't hold your breath waiting for 60%. And don't wait for 60% before modifying precautionary measures. That may never happen.

...

Community-wide serology tests will also help doctors and researchers learn more about the body’s immunological response to the novel coronavirus, especially about how long immunity lasts and what specific antibody level conveys immunity. UM’s testing only detects the presence of antibodies and does not measure the level, Kobetz said.

This last bit is very important. The presence of antibodies does not--repeat, DOES NOT--serve as a simple signal the presence of effective immunity or even any degree of immunity. Nor does it tell us how long that immunity lasts. The common cold--a coronavirus--confers immunity that lasts for only two months. There has never been a vaccine developed for any coronavirus.

UM researchers used statistical methods to account for the limitations of the antibody test, which is known to generate some false positive results. The researchers say they are 95% certain that the true amount of infection lies between 4.4% and 7.9% of the population, with 6% representing the best estimate.

...

The results are similar but not identical to other serological surveys in California, but indicate far less infections than a recent survey in New York, which found a nearly 14% infection rate statewide and upwards of 20% in New York City.

Mina, the Harvard epidemiologist, said infection rates based on tests done when the virus is active — the most common form of testing to date — likely under-represent the true extent of the pandemic about ten-fold.

But Mina cautioned that the infection rate is likely to be higher in densely populated communities than rural areas. “This is not a virus that has homogeneous spread,” he said. “This is a virus that has clusters of really, really high infection rates and then there will be areas where it’s just not so much.”

What that means is that there's much more at work here than what most models have assumed. The virus is unlikely to have a "really, really high infection rate" in one place and a low one elsewhere. Different rates of infection from one location to another are more likely accounted for by factors involving the closeness of human interactions in given areas, the likelihood of receiving a significant "dose," as well as the susceptibility of the given population group that is exposed. This is similar to what I've been saying about "human geography," and since it directly contradicts the claims of most modelers that the virus will spread everywhere at a relatively rapid rate, scientists and public health policy makers need to focus on this.

UM researchers say their findings are more robust than most because they used Florida Power & Light to generate phone numbers in targeted demographic areas, leading to a more randomized selection of participants.

Results from a serological survey in Santa Clara County, California, released last week, estimated between 2.5% and 4.2% of the population was infected, but the survey came under widespread criticism for, among other limitations, recruiting participants through Facebook ads, which scientists say could skew results by including people who think they got sick and want to confirm their suspicions.

That survey found an infection rate that would be 50 to 85 times higher than the official number of confirmed cases, while UM’s more randomized study indicates an infection rate about 15 times higher than the official case count.

Natalie Dean, a biostatistician and assistant professor at the University of Florida, has been following serological surveys across the country. She said UM’s preliminary results show a level of under-detection that is roughly consistent with the New York results and other estimates — aside from the more controversial Santa Clara study.

“I would expect this factor to be somewhat stable across the U.S., although it will vary due to differences in testing capacity,” Dean said.

...