OK, I'd like to raise the point and ask questions about what I think are the two possible main strategies for dealing with an epidemic:
① stop it at all costs,
② slow it down but let it run its course until immunity makes it stabilizable.
•1/32
① stop it at all costs,
② slow it down but let it run its course until immunity makes it stabilizable.
•1/32
In a given population, the infection has a certain reproduction number. If that number is >1, it will tend to grow exponentially. Exponential growth cannot go on forever in a finite population: the reproduction number has to drop. But what causes it to? •2/32
Well, to lower the reproduction number, our two weapons are: how we behave, and our immunity. (Or death. Our three weapons are behavior, immunity, death; and ruthless efficiency… Never mind.) Let's assume people who were infected are made immune from the infection: … •3/32
… so people who have recovered are effectively removed from the population, not just as potential for infection but also as vectors of infection. Thus, the number of person-to-person links that the infection can use to transmit decreases as more and more recover. •4/32
After sufficiently many people have recovered, even absent any particular measures to contain the infection, the effective reproduction number drops below 1 and it eventually dies down. And this is stable: if more cases appear, they won't be able to progress. •5/32
This is a phenomenon known as herd immunity: if a sufficient proportion of the population is immune from an infection, they protect the entire population by essentially disconnecting the links through which the epidemic could propagate (percolate). •6/32
How many people need to be immune for this to happen? Well, in the ultra-simplistic model of a homogeneous population with random contacts, 1 − 1/R₀ where R₀ is the basic reproduction number. See: •7/32
This would predict ~70% for R₀~3. But this model is too simplistic: observation suggests that true figures are much below this. And indeed, social structure is not random: an epidemic will hit highly connected individuals first, … •8/32
… and by making highly connected individuals first, the epidemic starves itself of its own fire. We don't know how many people actually need to be immune for the epidemic to be stable, and it depends on the disease, the type of transmission, social structure, etc. •9/32
But it remains the case that there is a critical fraction of immunity beyond which the infection can no longer spread, and it can be effectively contained. Typical ballpark values might be around 20%, maybe more, maybe lower, but certainly not below a few percent. •10/32
(Note that I'm not talking about the percentage of people who will eventually be infected if we do absolutely nothing to stop the epidemic. That number is way higher, see — I'm talking about the fraction at which it can be stably contained.) •11/32
So long as the critical immunity threshold hasn't been reached, containing an epidemic is tremendously difficult. We can do it by lowering the effective reproduction number by other means, namely, changing our behavior: how many people we see, how we wash, etc. •12/32
All these things will change the reproduction of the epidemic (if everyone wears a hazmat suit all of the time, the epidemic cannot propagate! if we're just a bit more careful, it slows down a bit, etc.) Quarantines and isolation measures are of this kind. •13/32
So, what are the two strategies I mentioned earlier? Well number ① is: don't rely on immunity, change our behavior sufficiently (including by closing public places, forbidding all gatherings, etc.) so the reproduction number drops below 1 and the epidemic dies out. •14/32
The problem with this is that it's unstable. If you contain the epidemic that way while a still-negligible proportion of the population has been infected, there is no immunity to prevent the epidemic from flaring back to life in an exponential growth. •15/32
So you need to make the changes permanent, or at least ready to be deployed at any moment until every single last one case of the infection has vanished from the face of the Earth, something which is very hard to know. •16/32
Strategy number ② is a compromise between number ① and “just wait until immunity stops the disease naturally”. The idea is to let the epidemic run its course, not stop it, but try to manage it: most importantly, SLOW it down, or, as they say, #FlattenTheCurve. •17/32
The point is not to stop the epidemic entirely as in strategy ① but change our behaviors just enough to make it progress at a speed which makes it manageable by the community's health system, even if the total number stays the same. •18/32
(Also, try to control who gets infected: protect the most vulnerable people most carefully, so they don't get infected ever, and relax those controls when enough less-vulnerable people have been infected and herd immunity makes it less risky.) •19/32
So, those are the two basic strategies: ① strike strong against the epidemic and stop it completely but be aware that this is unstable, or ② strike just enough to make it manageable and wait until immunity protects in a stable way. •20/32
Of course, it's not an all-or-nothing choice. If the epidemic gets out of hand, both strategies become identical: do everything we can to prevent contagion. But still a choice has to be made as to how hard we try to contain the infection. •21/32
But which is best in the case of #Covid19? Now here's where it gets tricky, and I really don't know, and I think there should be an honest debate about this which is why I'm laying the options as clearly as I can. •22/32
Strategy ① got us out of SARS (version 1) which was much deadlier than Covid-19. But it had characteristics (in terms of symptoms, recognizability, incubation period, etc.) which made it easier to track down. Still, a MAJOR disaster was avoided that way. •23/32
China clearly chose option ① in dealing with Covid-19. So, it's possible. South Korea seems to be showing that it's possible even in a democratic society. But the problem remains with option ① that containment measures need to be permanent, … •24/32
… or at least remain an option on the book for the foreseeable future, because the situation is unstable. What is the cost of closing all schools and all public spaces, severely restricting travel, etc., for years? This is a difficult question. •25/32
So maybe option ② is preferable? (Or maybe we just don't have a choice and fail at doing ①.) Well, for option ② to succeed, a few percent of the population at the very least have to be infected. •26/32
I thought this was clearly the best route: just #FlattenTheCurve to make the epidemic manageable, until maybe 20% (←random guess) of the population have become infected and then things are more stable and the epidemic can be halted without drastic measures. •27/32
But now here's where I'm really terrified: Italy currently has ~0.015% of the population reported as infected, and the health system is already struggling to cope with the load. Less than ONE FIFTIETH OF ONE PERCENT! •28/32
OK, maybe cases are underreported and maybe the overload only concerns the regions with the highest rate of infection. Even then, it can hardly be above maybe 0.5% infection rate. I have a hard time wrapping my head around how things can already be so grim! •29/32
This makes it very difficult to picture how we can proceed with strategy ② without a major catastrophe: how do we get to maybe 20% people having been infected when 0.015% seems already unmanageable‽ •30/32
I don't know. Both options seem terribly unappealing, and I don't know what to say. But I think experts should really discuss these two strategies and politicians should make it clear what they're trying to achieve. •31/32
If ①, then how can we possibly relax the lockdown without the infection reappearing every time we do? And if ②, how can we possibly manage to reach 20% (or whatever) infected in a manageable fashion? Do you vote for Scylla or Charybdus? Which is the lesser evil? •32/32
Clarification: there are estimates of per-country reproduction numbers R on the wonderful cmmid.github.io/topics/covid19…
① is → implement severe lockdown until epidemic has disappeared from the Earth;
② is → lockdown just so as to decrease R to ~1, not much more.
•33/(32+1)
① is → implement severe lockdown until epidemic has disappeared from the Earth;
② is → lockdown just so as to decrease R to ~1, not much more.
•33/(32+1)
