Thread by @toxvaerd1

The Zero COVID strategy is gaining ground, with countries like New Zealand standing out for their success in dealing with the epidemic. Here are some observations on why and how this strategy works. A thread. #EconTwitter #EpiTwitter #eradication #ZeroCovid

The main ideas are borrowed from previous work I've done in slightly different context, looking at how disease control with multiple policies should be designed. There I looked at treatment and prevention, but ideas similar.
inet.econ.cam.ac.uk/research-paper…

Suppose we want to control disease and have two broad policy instruments available, lockdowns and test-trace-isolate (TTI). Each policy is costly, so we need to understand how to use and combine them wisely at different phases of epidemic.

Because interventions are costly, we want to get as much bang for the buck when we use them. The thing is, the value of each intervention depends on how much infection there is in the population. So effectiveness of policies changes over time. Let's consider each in isolation.

The value of lockdowns *increases* as infections surge. This means that we want to keep people at home when things are getting out of hand. For low prevalence, the marginal benefit of lockdowns is lower than marginal cost, and other way around for high prevalence.

This is roughly what most governments do. They lock down for high infection numbers and relax when infections fall. Problem is: by using *only* lockdowns, we cannot generally hope to eradicate disease. See thread on why this is so here:

https://twitter.com/toxvaerd1/status/1323712484583329793

In turn, the value of test-trace-isolate *decreases* with prevalence. The marginal benefit of TTI is higher than marginal cost when disease prevalence is low but lower when infections soar.

Let's dwell on this last point. Note that the more effective TTI we do, the more we manage to suppress new infections. This helps decrease prevalence and hence make TTI *even more valuable* and prompting us to do more if it. There are dynamic complementarities.

Let's return to the Zero COVID strategy and the alternative that relies on only lockdowns. With lockdowns and no workable TTI, the best we can hope for is to follow the underlying dynamics of disease but to attenuate effects. This is worthwhile but we can do even better.

With strong lockdowns and effective TTI, we can get numbers down and *keep them down* at relatively modest cost. Once we've "invested" with costly lockdowns, we can reap the benefits and maintain low prevalence at relatively modest cost through TTI.

Note that we cannot rely on TTI when infection numbers are too high. Tracing infections is unwieldy and exorbitantly expensive when many people are infected. But it's doable when infection numbers are low, as we have seen in many countries that have done it successfully.

Not getting test-trace-isolate infrastructure working is a serious mistake, because it robs government of a key instrument in managing the epidemic. Without this tool, Zero COVID is completely unrealistic.

The UK government may have started to realise this, having introduced "surge testing" to stem the spread of the South African variant.
gov.uk/government/new…

But once the current lockdown brings infections down to manageable levels, the government should treat *any* COVID infection in the same way. Of course, this is only possible with a working TTI infrastructure and building this should now be the primary policy objective.

Caveat: the linked paper looks at "treatment", not TTI. In that work, treatment works by taking infected people out of population and thus reduces disease incidence. TTI works in similar fashion, by reducing transmission. Check paper out for details.

A few shout-outs to researchers who have made similar points for some time: @devisridhar @_Maia_King @sjwrenlewis @t0nyyates

The tradeoffs I described above are also at the core of a forthcoming paper by
@alvafer64 @dargente05
and Francesco Lippi. Check it out! aeaweb.org/articles?id=10…