The CDC’s description of how COVID-19 spreads is incomplete. It simply does not fit the facts. That began to bother me more than a month ago, and I summarized the inconsistencies in Post #573, 3/26/2020.
I’m now think I have an understanding of this disease that fits all the facts known at this time. The point of this post is to provide that accurate description, and highlight what you need to do, in addition to what the CDC says, to protect yourself and help contain the spread of coronavirus.
I did not put in citations as to sources here. I’ll revise this later in the day to do that. But I can back up everything I say in this post.
And this is the last I will write, along these lines, because I’ve finally satisfied myself that I have an explanation that fits the facts. I’ll only revisit this if the facts change materially.
The upshot is: Do everything the CDC tells you to do. And more.
What the CDC said originally.
The CDC initially said that this disease was spread primarily through droplet transmission by symptomatic individuals. That is, somebody who knows they are sick coughs or sneezes nearby. When that happens, the emit a cloud of “large” droplets that act like little projectiles, and spray everything within six feet or so. If such a droplet lands in your mouth, nose, or eye, you can be infected. If you touch a surface where such droplets landed, and then touch mouth, nose, or eye, you can be infected.
“Large” does not mean visible to the eye. Large, in this case means greater than 5 microns. The thinnest human hair is about 40 microns. These are “large” only in the sense that they will (normally) quickly settle out of the air.
And the virus eventually self-destructs on surfaces. So the longer those droplets sit, the less danger of infection. You see a lot of scare-mongering over this, so you should not believe every bit of click-bait thrown your way. If you can’t track the source back to a reputable scholarly publication (NEJM, JAMA, The Lancet, Science, Nature or similar), you’d best disregard it.
This description of disease spread is the rationale for the social distancing rules. And for washing your hands at every opportunity. And for avoiding touching your face. You should continue to do all of those.
This is also the reason they did NOT recommend wearing masks in public unless you thought you were sick. In theory, social distancing was enough to prevent droplet spread. And, in theory, you weren’t infectious until you started to feel sick. So, the theory went, masks weren’t necessary.
The CDC was, I believe, largely guided by the World Health Organization report on the Chinese experience (Post 551, 3/16/2020). And as it turns out, some aspects of that report were either wrong, or do not apply to the US experience with this disease.
Finally, I have a fear that I cannot possibly confirm, which is that the CDC stuck with that view for so long by confusing absence of evidence with evidence of absence. Something like 96% of all cases in the US are still “under investigation”. I can’t tell whether that’s because they haven’t gotten to them, or whether they looked, and if they didn’t find the source, they keep the case open.
For sure, when they did contact tracing and they found other sick individuals, they could plausibly attribute the spread to droplet transmission/close personal contact. And so, they closed the case. But it’s not clear how they classify cases where they did contact tracing, and found no such individuals. For sure, there’s no category for them in the CDC’s public-facing information (below). On the CDC’s contact-tracing summary page, there’s no category for “community spread” cases. Their counts show cases where they identified the source, and cases that are under investigation. There’s no count of “we looked and couldn’t find the source”. At least, that’s the plain-language reading of what’s on their website.
By contrast, if you look at the news reports, in many cases a “community spread” case was typically identified among the first 5, 10, or at most 20 cases in a region. A “community spread” cases is one where they tried to identify the source of the infection but were unable to do so. I don’t think I can infer from those newspaper reports what the rate of community spread cases actually was, at the outset of this epidemic. But there were enough of them to suggest that, at that time, a) people got coughed and/or sneezed on by a stranger and didn’t notice, or b) some other mode of transmission was already operating. And if it’s the latter, and we have effectively shut down most close contact/droplet transmission (outside of the home) with social distancing and so on, then whatever-it-was that was generating those community spread cases early on is, arguably, a far more important means of disease transmission now.
What the CDC says now.
The CDC made two subtle but critical additions to their description of how this disease spread.
First, they changed “cough or sneeze” to “cough, sneeze, or talk”. So now, they acknowledge that you emit drops (of some sort) even when you merely talk.
Second, to the phrasing about “land in your mouth, nose, or eyes” they added “or inhale”. So now, they acknowledge that these drops might be of the type that you are likely to inhale.
And the CDC changed its guidance to add “wear a cloth mask in public”.
The changes to the description only make sense if the CDC has come to realize that COVID-19 commonly undergoes aerosol transmission. That is, it can be spread by very small droplets, of the sort that stay suspended in the air for long periods of time, and will drift on wind currents. If you haven’t done this already, I urge you to go read the article cited at the top of the front page of this website. Just look at the diagrams, and you’ll get the picture.
But the CDC will not openly say that aerosol (airborne) transmission is significant path for spread of disease in the general public. I’m not sure why. Arguably, they don’t want to panic people. Arguably, they don’t want people scrambling for N95 respirators (presumably in some fictional universe where citizens actually could still buy those in large numbers.) Possibly, the CDC’s position is what it is because they have evidence that it isn’t happening in large numbers, and for whatever reason, they will not release that.
My best guess, having worked with government health agencies for much of my career, is that the CDC literally can’t change its description of the disease in this way. That’s because a) they don’t have high-quality evidence that it is happening in large numbers, b) they probably never will have high-quality evidence this is occurring, and c) they won’t change their public-facing information without that high-quality evidence. In other words, they likely have a written policy dictating that public-facing statements of this type can only be made if there is a high-quality evidence basis for doing so. So their public-facing information remains in this hedge-your-bets position where it only makes sense if aerosol transmission is likely, but the CDC will not openly say that.
Why won’t they ever have evidence of airborne transmission? Even for disease that are know to have airborne spread, the only way the CDC knows that is to infer it from certain unique, well-studied events. Events where the disease was clearly documented to have traveled (e.g.) between floors of hospital via a poorly-designed ventilation system. So it’s not like the CDC is going to get its carefully researched, correctly documented information on this disease, in real time.
That said, their advice to hospitals, and to the US Army Corps of Engineers, clearly acknowledges that this is an issue in the hospital setting. That’s why this disease calls for N95 respirators. That’s why the Army Corps of Engineers produced plans for converting hotel rooms into expedient “negative pressure” isolation rooms that would keep aerosolized coronavirus from spreading from patient rooms out into common areas. But there, you have very sick individuals (high virus shed rate) living in a small enclosed space (hospital room) and others must be exposed to that air for lengthy periods of time (hospital workers.)
Finally, the new advice on masks is one part public health, one part social engineering.
The public health aspect is, wear a mask in public. If you were solely concerned about the health of the public, you’d stop there. In fact, you’d recommend that people wear the best mask they could obtain.
The social engineering part is “cloth mask”. Homemade or store-bought fabric masks are inferior to hospital-grade equipment, including a properly worn surgical mask. The CDC simply doesn’t want the public to buy such masks, as they are in short supply for hospital workers. And apparently the CDC lives in a fictional world where the public could still buy such masks in large quantity. (But, in fact, all significant retail outlets are tightly controlled now, and no responsible company (e.g., Home Depot, Amazon) will sell such masks to the public. You can probably still roll the dice on Ebay, but that’s not a channel for a significant number of masks.
The CDC’s description does not match the facts.
The CDC’s description of COVID-19 transmission simply does not explain the facts. Not even close. Not even initially. And certainly not now.
First, I’m not going to document the particular posts, but people have not been coughing or sneezing in public for weeks now. I won’t say its a never-event, but it’s close. Not-being-sick-while-shopping is now a rigidly enforced social norm. And anyone who feels sick knows enough not to be out in public. (Work, though, is a different issue, discussed below).
Second, we continue to see famous people, and world leaders, getting sick with this. It’s beyond implausible that the British government let somebody who was coughing and sneezing anywhere near their Prime Minister. Ditto, major networks letting the same near their star newscasters and commentators. And so on.
Third, numerical modeling of the Wuhan epidemic strongly suggested that the number of people who were spreading coronavirus was far larger than the number documented through testing. They literally could not make the math work without factoring in a population of infectious individuals that was several times larger than the know, tested, documented population. They may have overestimated the number, for reasons I discussed in an earlier post. But, a large number, I think that’s reasonable to say.
Given that our epidemic is spreading faster and to a greater extent than the Chinese epidemic, it’s almost a certainty that this must be true here as well. That if you did the math, on the spread of epidemics, you could not possibly explain it based on the cases that have been positively identified. (Unless you assumed that this is wildly contagious, far more contagious than anyone thinks. And I believe that’s a far less likely scenario.
Fourth, we now have concrete evidence that, among a large population of pregnant women at one New York City hospital, the vast majority of COVID-19 carriers are asymptomatic. From that, it’s a pretty good guess that this true of the younger population in general. And somewhat less good a guess, of the population in general. Not only did they not have symptoms at the time, the vast majority of those never showed symptoms for the entire period they were under the care of that hospital system.
The only thing we do not know for sure, about that asymptomatic COVID-19 population, is how infectious they are. How capable are they of spreading the disease? Traditionally, for many viral disease, the degree to which a person is “shedding” virus corresponds well to the severity of the illness. If you look sick and feel sick, that’s probably when you’re shedding virus into the environment at the fastest rate. Here, it’s not clear how much virus is shed by asymptomatic individuals. And we don’t know if there are true “Typhoid Marys” in that population or not — individuals who are permanently that way — or whether these asymptomatic COVID-19 patients eventually rid themselves of the disease. By far, the more plausible guess is that there are no permanent carriers.
Fifth, I have heard first-hand description of one case where individuals, returning from overseas, spread COVID-19 in a workplace setting, before they developed symptoms. I have read many more. That’s why organizations that take this seriously are now, in some cases, simply quarantining individuals returning from overseas. So, for sure, between the time you are infected and the time you show symptoms, in at least some cases, you can spread the disease. The “pre-symptomatic” population is clearly infectious.
Sixth, there is at least one “superspreader” event that cannot possibly be explained by droplet transmission. That’s the Mount Vernon, Washington choir practice, where a single choir practice, run in a properly sanitary manner, resulted in 45 out of 60 individuals being infected with COVID-19. The only sensible explanation of that is aerosol (airborne) transmission of COVID-19.
And I’ll bet that this spate of high infection rates in meat processing facilities is the same, though I have no direct evidence of that. I believe I’ve now heard of two such incidence — one a pork-processing plant, one a poultry-processing plant. Those people have, as I understand it, fairly rigid discipline about hand sanitation. But, also, as I think I understand it, those are noisy (shout-to-be-heard), aerosol-rich environments, which would be the perfect place for aerosol transmission.
I mean, plenty of people still work in factories. What’s unique about meat-processing plants? So I’m on the looking for other workplace-related incidents like these. Looking for places that have the same type of conditions, but aren’t meat-packing plants.
A more accurate description
Here’s my best guess for an accurate description of the situation. One that explains the facts. This is based on everything that I’ve learned about this so far.
First, just as the CDC says, COVID-19 can be spread by droplet transmission by people who have symptoms of the disease. So, do everything the CDC tells you to.
Second, the number of individuals who are infected and capable of spreading the disease is far larger than the population with symptoms. Which is itself larger than the population that has been tested and known to have COVID-19.
Some people show no symptoms because they haven’t gotten sick yet. Those individuals are clearly capable of spreading the disease. Some people apparently never will have symptoms. It’s not clear whether or not those individuals are capable of spreading the disease.
Third, this disease is spread by short-distance aerosol transmission. Simply standing near someone and being talked-to puts you in the path of droplets, some of which will be aerosol-sized. The louder you talk, the more aerosols are produced.
Fourth, this disease may be spread by intermediate-distance aerosol transmission, which is the term I would use instead of airborne transmission. That is, you can in theory contract the disease just by sharing the same enclosed space with an infected-but-asymptomatic individual who talks, shouts, or sings in that space.
And, some individuals are aerosol super-emitters, producing an order-of-magnitude more aerosols than the average. Being in a room with one infected individual who is an aerosol superemitted, and talking, therefore generates the same airborne viral load as being in a room with (say) 20 sick, but non-super-emitting, individuals. Such aerosol superemitters are few, but they are not rare. My recollection (I’ll go look this up) is that it was low-single-digits per 100.
We hope that this is a rare transmission path. But we’ll never know. The only time this can be pinpointed as an infection path is during extreme events, such as the Mount Vernon, Washington choir practice event. Otherwise, you get indirect evidence based on swabbing down a few hospital rooms, beyond “droplet transmission” distance from a patient, and seeing if there is evidence that aerosols landed on those surfaces. But, as I said in an earlier post, I doubt those COVID-19 patients were singing.
Fifth, a lot of disease transmission is occurring in the workplace, rather than in public spaces. I say this not only because of the big events (e.g., the pork-processing plant with 300 sick out of a 3000 person workforce). I say this because, in Virginia, the documented prevalence of disease by age is hugely at odds with the documented risk of illness by age. Putting the oldest-old (likely nursing-home-spread) cases aside, the prevalence of this disease peaks in the working-age population, not in the retiree population. So my guess is, people are picking this up at work. You need to treat any indoor work place that you share with others using the same care you would treat any public space. If you wear a mask while in the grocery store, you should have the sense to wear one while at work.
This graph, below, based on Virginia Department of Health case counts, and Census population counts — this is what is grossly at odds with what we should see, if all of these age groups are equally exposed. The children and oldest-old make sense. The peak in the working-age population makes no sense at all, unless they are getting far, far more exposure than the retirement-age population.
Beyond the CDC guidance.
The description above does provide an adequate explanation of the facts. It explains why this disease continues to spread despite rigorously-adhered-to social distancing measures. Despite seeing no sick individuals in public. Despite having a generally educated population (in Fairfax County) that has the good sense to follow CDC sanitation rules. And despite large organizations taking pains not to allow sick people near their key personnel or star performers.
So here’s my advice, above and beyond what the CDC says. New rules.
- Any time you are in an enclosed space with others, other than your home, wear the best mask you can get. This means in public and at work. If you have one with any known capability to filter aerosols, use it.
- Keep your voice down in those spaces, and encourage others to do the same. The louder you are, the more aerosols you emit.
- Expand your social distance if you are walking, running, or biking behind someone, due to the slipstream effect.
- Don’t assume that a person who appears healthy cannot infect you. They can. If they have COVID-19 and haven’t developed symptoms yet, they definitely can. If they are one of the (apparently) numerous truly asymptomatic individuals, in all likelihood, they can.
- Don’t assume that because you feel healthy, you cannot infect others. You can. Same as rule 4: You might have it and have not developed symptoms yet, or you might be one of those (apparently) numerous truly asymptomatic cases. My guess is, such truly asymptomatic cases are far more prevalent among younger rather than older people.
Christopher Hogan, Ph.D. chogan@directresearch.com
