Thread by @joel_c_miller

Melbourne has gone back into a lockdown...

Some thoughts, explanation, and a warning for other parts of Australia that aren't (yet) in lockdown.

A few weeks ago we were at 8 cases a day, and we had been close to elimination.

That quickly increased to 20 cases, so some restrictions were introduced.

Unfortunately with COVID-19 the response has been plagued by several factors:

1) It takes 2-3 weeks typically for interventions to have an impact that we see in case counts.

2) If an intervention is too small, all that means is that the exponential growth slows down.

3) Once it starts to shrink, it does so very slowly. So the time it takes to double is much less than the time it takes to halve.

After an intervention it takes weeks before we know whether cases have gone from doubling say every 5 days to decreasing or to just doubling every 10 days.

If it's still growing a few weeks later, then it's a GROWING epidemic that's 10-20 times as large as it was.

Because it grows faster than it falls, that means that for every week you delay putting in place a stronger intervention, that means more than one additional week that intervention must be in place.

This asymmetry means that the main lesson of outbreaks elsewhere and here in Australia is this:

Act Fast.

Australia did a lighter lockdown than the UK. Australia almost eliminated infection. The UK still has many cases, with much worse lockdown impact.

Why did it work in Australia?

Australia did theirs when the epidemic was much smaller.

That's it. Because it was still small, a shorter, lighter intervention was much more successful.

This is why VIC re-introduced some restrictions when our daily count was still just 20, because the doubling trend was clear, and growth has continued for just over 2 weeks now.

We may soon see benefit from those early interventions. If we stablize this week, that would be success of those interventions.

But at hundreds of cases a day, the disease would continue to spread into parts of Melbourne without those restrictions.

So those restrictions would have to expand to track the spread.

Playing catchup to a disease that grows exponentially with lots of unseen cases doesn't work (see Europe and North/South America).

I am glad that regional VIC is not included in this lockdown, and it is fortunate that the rest of Australia seems to have escaped infection so far.

For these parts, I hope that seeing how quickly it grew in Melbourne serves as a warning.

You don't have to be as cautious as we do, but it is important that you maintain behaviors that will slow the spread.

Actions that slow the spread mean that if it reaches your region the growth may be slow enough that a mild response is successful.

I would personally encourage mask-wearing.

It's not perfect, but anything that slows the growth is important.

Lots of personal actions can slow down the spread. Assume that the people you interact with one day are infected. And then when you make a decision about who to interact with a few days later, assume you are now infected.

But from a policy point of view, the policy has to stay ahead of the infection. You have to assume that what you're seeing today is the situation 2 weeks ago.

In sports you defend where your opponent is going, not where the opponent is.

We must defend against where the pandemic is going to be next week. We can't defend against where it was 2 weeks ago.

So for those criticising the response as an over-reaction, or too widespread: you might be right.

But if they follow your advice and you're wrong, we won't know for 2 weeks. If the probability you're wrong is 10%, are you willing to risk a widespread epidemic on that?

So policymakers are facing terrible choices. An under-reaction for two weeks means more than 2 weeks of additional stronger response, or accepting a widespread epidemic. @TimSmithMP has made his choice clear.