Source of the post
Pretty interesting. Testing in Iceland, however, indicates that as many as 50% of those infected have no or hardly any symptoms. If the models had been run in such a configuration, I wonder how the results would change.
I think we can gain some intuition by looking at the models he runs starting at 6:39, which show the effect of changing the fraction of cases that are quarantined (e.g. because they show no or weak symptoms and don't get tested). In this model, 100% of cases being quarantined is very effective at stopping new infections, while if 80% are quarantined then the curve is flattened but with a long tail, resulting in twice as many total cases.
So having some fraction of cases not being detected and quarantined causes a disproportionately greater number of people catching the virus. If about 50% of cases are nearly asymptomatic and not known, then measures to slow the spread by identifying and quarantining infected by symptoms are dramatically less effective (see for example the simulation with communities at 50% quarantine at 8:35 -- it's only barely better than if no action was taken at all!) This goes with his insight that it is not the diseases that have the highest mortality rates that are the most dangerous to the world. It is the ones that are lethal to some, but weakly symptomatic or even asymptomatic in others, which makes it more difficult to track and contain its spread.
How to solve this problem? The most obvious solution is to test everybody. But in many countries this is not so easy because testing supplies are limited. The rate of testing capability must increase faster than the rate of new cases.
Another way is through contact tracing. If we can trace back every person a confirmed infected person was in touch with, such as through cell phone data, and get those contacts to all self quarantine, then that dramatically helps to slow the spread because we catch many more people before they develop symptoms, or even if they would not show symptoms but still spread it. This is how Singapore acted, and now they're wanting to make the technology more widely available. I can easily imagine a lot of people not wanting to use such measures for fear of invasion of privacy, but I think for the amount of good it can do it would be a wise option. Perhaps the fears could also be culled somewhat by ensuring oversight by an independent organization, and that the use of the technology goes away once the pandemic ends.
Source of the post
And one thing these simulations miss is the mortality distribution across the population. If a disease is perfectly harmless for people under, say, 70, and mortality is very high for those above, then, if the disease has already spread, effective isolation of the vulnerable while the rest quickly become "recovered" would be very effective.
I find this very sensible, though I also fear a few potential problems when applied to the real situation.First is the size and complexity of the vulnerable population. It seems quite clear that mortality rates increase dramatically at higher ages, so it is paramount to protect them the most. But it is also not simply zero below the age of 70, or even 50. As we've seen, the mortality rate in all age groups is about 20 times higher than the seasonal flu, and under the age of 30 it amounts to about 1 in 500 cases.Now we may expect the true mortality rate is less than that because the fraction of people in those ages who show symptoms and get tested is also smaller, but I don't think it can be much less. I could believe a factor of 2 or so, but probably not a factor of 10, and that also doesn't seem to be borne out much from the data we're seeing from outbreaks in cruise ships or other areas where testing is more robust.
Add to this the effect of people with other health problems that increase their risk such as asthma, diabetes, and hypertension. So it's really a large number of people across a broader spectrum than age that will want to avoid getting this virus at all, and I worry that if we tried to protect these who are most at risk while letting the virus spread through the rest of the population, it could prove more difficult and we may still see an alarming number of deaths. I very much hope I'm wrong, but I wouldn't want to gamble lives on it.
My second concern is that if the virus is allowed to spread through much of the population quickly, there are that many more chances for a new strain to diverge which is more lethal to different age groups, complicating mitigation efforts further. We know that this is not impossible. In fact it happened in the 1918 pandemic, where the second wave of infections was far more deadly, especially to those aged 20 to 40. That was perhaps an unfortunate case of artificial selection, where the soldiers at war who had the milder strain stayed on the battlefield, while those with the deadlier strain were sent home, enabling it to spread further.
SARS-CoV-2 is a new virus in humans, and we're learning more about it, and how to best respond to it, very rapidly. My hope above all else is that we can minimize the amount of deaths without hurting economies too much in the process, and I do not see these goals as mutually exclusive.