Every conceivable Covid-19 outcome

The reality of the Covid-19 pandemic's staying power has begun to sink in. As surreal as it may seem, we are starting to admit to ourselves that we might be stuck with self-isolation and social distancing for a long time.

There's still more than enough uncertainty to make anybody cautious about making predictions about how and when the world will return to normal. My sense is that trying to estimate when the crisis will end is a bit silly. But what is especially interesting and scary to me is that we still aren't quite sure what the endgame looks like.

It does seem like lately a lot of governments and institutions have converged around a strategy. I first became familiar with this strategy in the Medium post "The Hammer and the Dance", and it goes like this:

  • Enforce heavy social distancing via lock-downs. Halt the spread of the virus until infection numbers have dropped dramatically. Meanwhile, quickly scale containment tools such as testing and contact tracing.

  • Then, begin to gradually release the measures, so that people can slowly get back to something somewhat resembling a normal life and the economy can restart.

  • Continue in this semi-normal state until either a vaccine has been mass-produced or treatment options become radically better.

I have no intention of being cynical; this outcome would be quite positive. According to the Gates Foundation, it would mean that within 2 months or so, people in most developed countries would be able to go back to a life that kind of resembles the one we had. Sure, we won't be going to concerts or working without social distancing protocols in place any time soon, but it would be vastly less soul-crushing than being locked up in our homes.

However, it wasn't too long ago that the words "herd immunity" and "flatten the curve" were all you'd hear when discussing the endgame. My point being that as we acquire more information, our plans could change. More explicitly: we don't really know if we will able to successfully release measures (enough of them to go back to a semi-normal life) without the virus spreading out of control again. If so, our plans could very well change again.

That's why I thought it would be an interesting experiment to try to write down every possible way the crisis could unfold. What follows is my attempt to draw a flow chart that maps out the different decisions our leaders can choose from, and the chain of events that could follow. If you notice any gaps in my thinking, I'm open to suggestions about how to modify the flowchart to cover them.

What the flowchart will let us do is to start thinking about the endgame probabilistically. We can think about how likely each of the paths looks like with our current understanding of the crisis. And we can update this probabilities as we continue to gather information and the uncertainty lessens.



The hammer and the dance path seems to be the strategy most countries seem to be moving towards currently.

In this path, the easing up of measures is successfully implemented without a second wave of infections happening. The thinking here is that we weren't sufficiently prepared to contain the virus during its initial outbreak, but that we will soon have learned and prepared enough to be able to drastically deescalate measures while keeping the virus contained.

The exact way in which this would be done will vary from country to country, but some of the strategies I've seen floated are:

  • Drastically scale up testing.

  • Faster testing (<24hs to obtain a result).

  • Drastically scale up contract tracing. I've seen estimates that 100k contact tracers will be needed in the US.

  • Produce a lot of masks, and force people to use them when they are out in public, especially on crowded spaces like grocery stores.

  • Opening up must be done gradually, in order to have enough time to assess whether lifting/changing any given measure does not increase infection rates massively. Spain is moving to a deescalation phase in which measures will be lightened in 2 week intervals to ensure the government can sufficiently evaluate whether infection rates increased.

  • Figuring out protocols for reopening places without increasing infection rate significantly. Some examples would be planes without a middle seat passenger, restaurants with every other table empty, workplaces with only 50% of the people there and physical distancing between them, etc.

  • Using apps and credit card history to do better contact tracing. Google and Apple have released an app for this exactly.

  • Allowing movement within more granular areas, say at the district or city level instead of at the country or province level. Mathematicians at Oxford have proposed a Green Zone model for this. Facebook has released a map of symptoms by county.

Some interesting reads on the subject are this medium article, this Gates foundation article and Spain's deescalation plan.

It remains to be seen if this path will work. It certainly seems to be possible in some cases, as countries such as South Korea, Taiwan and Vietnam have managed to contain the spread without enforcing drastic social distancing measures.

I would expect that measures can be deescalated to at least some degree, although the possibility exists that measures can't be lightened enough to give us anything close to a normal life. Another possibility would be that the virus is seasonal enough where the northern hemisphere can successfully deescalate until the end of summer, but will need to ramp up measures again on fall and winter.

The herd immunity path used to look more likely. Back when countries weren't taking the threat seriously enough to consider lockdowns possible (drinking game idea: reread February articles on the virus and drink every time Chinese lockdowns are called "draconian").

It fell out of favor in places like the UK once enough modelling was done to realize the death toll would be in the hundreds of thousands.

But it's entirely possible we could find ourselves again in a position where herd immunity becomes the only possible path.

If we were in a position where

  • Countries attempt the containment approach and realize there simply is no way to bring society back to a semi-normal state without the virus spiking

  • AND a highly-effective vaccine or treatment isn't close to mass production

it would become the most reasonable approach. Even if those conditions weren't fully met, at some point the death toll and lasting health effects of allowing the population to be infected will be more tolerable than the amount of human flourishing lost by maintaining a lockdown and the damage caused by its economic consequences.

In any case, at some point countries might have to switch strategies towards a herd immunity strategy.

Some open questions for me in this case are:

  • Could herd immunity be achieved in the short term without hospitals getting saturated? I haven't done the math myself. I've read articles that say it's impossible. However, Sweden seems to have decided to take this path from the get-go. We'll see how it pans out.

  • How much does the case fatality rate increase if hospitals and most patients have to undergo the disease without treatment? Maybe the IFR would increase, but not be orders of magnitude higher. After all, our treatment options are extremely limited, and even ventilators aren't very effective.

  • How much would the IFR decrease if countries released measures but mandate that at-risk people should stay isolated? Looking at statistics from Italy, only about 5% of deaths were from people under 60.

  • How common and serious are lasting health consequences for patients who recover from Covid-19? Even if strategies like isolating at-risk people reduced the death toll dramatically, the strategy could be unconscionable if it left a significant percentage of younger people with serious health issues.

And of course, the mother of all questions is: Would herd immunity work? Does the body generate a strong enough immune response to prevent reinfections in the short and long term? Are reinfections milder or more severe than the original infection? What would be the implications of the virus mutating significantly?

Metaculus predicts reinfections will be rare. My sense of it, from what I've read, is that most likely the cases of people testing positive after being "cured" have more benign explanations. The WHO however has warned that more research is needed to be sure that the herd immunity approach would work.