Thread by @DellAnnaLuca

DO LOCKDOWNS WORK? HOW MUCH? WHEN?

Lockdowns work in the measure people stay home so don't infect / get infected outside.

and do not work in the measure people still go out or still get infected at home.

How do these effects weight on each other?

(thread, 1/N)

Note: this thread is complex, please don't take any single tweet out of context and read the whole thing before commenting. Thank you.

2/N

The two most important parameters of a lockdown are:
- How many people keep going out and meet others.
- How many people are infected in a given area (neighborhood, etc).

The first parameter is intuitive: the more porous the lockdown, the less it's effective.

[3/N]

The second parameter is tricky, because as @michelteivel noted, people can get infected at home if they live with someone who carries the virus and, the more time they spend at home, the higher the chances of this happening.

[4/N]

See pic 👇: probs of contracting the virus function of time spent with an infected

(A) is nonzero because even without a lockdown people would still spend time at home.

(B) is lower than 100% because even if you spend a day home with an infected, you might not get it.

[5/N]

Now, the tricky part. The chart shows probability of contagion *conditional to living with an infected*.

This depends on how many people are infected in the area the people living in the house usually spend time in.

[6/N]

7/ As a consequence, counterintuitively, the short-term effectiveness of lockdowns decreases as the percentage of people infected rises.

Note "short-term" above.

8/ The effectiveness of lockdowns depends on:
- which % of the population is infected
- what the lockdown is for

In a little-infected population, the lockdown can produce quick effects.

In a very-infected one, the lockdown will keep producing new cases from in-family contagion.

9/ Lockdowns should in any case be an intermediate step used to slow down not total contagion, but contagion across groups (across cities, neighborhoods, families), so that the healthy groups can be separated from the sick; the former let back to work, the latter further isolated

10/ The effectiveness of lockdowns increases dramatically with testing.

Without testing, a family needs to remain fully locked down (fully, not even go out for groceries) for 3 weeks to be declared not infected.

With testing, this can be accelerated.

11/ Moreover, testing allows to isolate infected family members before they get to infect their close ones.

12/ The point of a lockdown is to isolate now in order not to isolate longer later.

Testing can reduce the time of isolation needed now.

Proper isolation reduces the time of isolation needed later.

13/ The point of this thread is not to quantify the effectiveness of a lockdown, but to illustrate the parameters that can influence it

Countries should not "cargo cult" lockdowns because that's what other countries are doing

They should use them with purpose, a means to an end

14/ Not all lockdowns are equal. We can have costly ones, effective ones, wasteful ones.

It depends from when they're implemented and how strictly and the relationship between the two.

It depends on whether testing is used.

It depends on communication on risks and outcomes.

15/ By the last point I mean that, depending on the percentage of population infected, the government should warn about the risks of going out (we'll have to stay home until we'll have learnt to stay home) & of staying home with an infected (anyone w/ symptoms, stay in your room)

16/ Summing it up: lockdowns are useful when used together with other tools such as testing.

They might be effective if carried on alone but early enough.

The might be costly if cargo-cult copied as the only measure, with the hope that they'll be enough to work.