I wanted to share a post by a friend, Joshua Goodman, which makes some great points. He articulates well that there’s a plan, and why it’s important that we distance down to low levels. Skeptics like to describe a false choice of (1) open up or (2) be closed forever. Between those, maybe opening up is better. But there’s a point to closing down; to get to where you can reasonably detect the small number of cases and contact trace them. Then you can reopen and slam the hammer down on cases before they spread with only 10,000 contact tracers in a state and hundreds of thousands of test kits. This is where Washington is trying to get – but your case count has to be really low.
If you let it get out of control it then exceeds your testing capability. But Georgia and Texas are opening up when there are already too many cases to trace! So that option doesn’t exist and they are going back to exponential growth.
With that, from Joshua:
I think everyone is wondering the same thing, namely, should we be reopening, and if not now, then when. And the answer is different in different places. I’ll talk about Washington State (we’re dong well – hang in there!), the country overall (flat – which is bad), and a couple of interesting states, Texas and Georgia (going to get a lot worse soon).
So, first of all, Washington State is doing surprisingly well. Over the past 7 days, we’ve had about 49 deaths, which is down from a peak of 167 deaths per 7 days, back on April 7th. It’s also down from 86 for the week before that – almost half. As I’ve mentioned, I think death rate and hospitalizations are the best way to measure progress – positive tests at this point tells us a lot more about our test capacity than how many infections we have. The hospitalization data is similar to the death data, down from a peak of 235 the week ended 3/15 to 80 (week ended 4/26), and this is on top of a background rate of very roughly 50 per week. (The hospitalization data is based on symptoms, not tests, and even before COVID-19 there were about 50 a week.) So, the hospitalization data suggests we are making even more progress, which is what you’d expect, since people are hospitalized a couple of weeks before they die, so when things are improving, hospitalization should be ahead of deaths. We’re also finally seeing an uptick in test capacity, which had been flat at about 3000 per day for a long time. (This is why positive tests have gone up even while other numbers go down.) In other words, all of this is great news – Washington State is making excellent progress.
The plan to reopen is when we get the total number of infections low enough, and the total number of tests high enough, that we can primarily rely on test+trace — finding infected people before they can infect others. This is what South Korea has done, and what has allowed them to stay open almost normally. In Washington State, we’re getting close to being able to do that – we need to keep it up a while longer, but we’re getting there.
If we open up too soon, infections rise, and it gets harder and harder to do enough testing, and we can’t transition to test+trace. If that happens, we either can’t reopen; or we can reopen with massive deaths — and everyone avoiding restaurants, travel, etc. – defeating the whole point. Staying closed a little longer will lead to a bit more short term pain, and a much better long term. It is a win by every criteria, both health and economics”).
OK, so Washington State is doing well. What about everywhere else in the USA? Well, on average, not so much. We had been hitting about 14,000 deaths per week, and now we’re down to about 12,500 deaths per week – progress, but very slow progress. Unfortunately, that news is rosier than it sounds. All of that improvement is due to progress in New York, which peaked at about 5000 deaths per week, and is now down to 2500. If you graph the US, with New York excluded, what you see is a rapid rise, and then flat for the last two weeks – neither up nor down. Flat is bad – at current rates, it’s going to be difficult to do enough tests at the national level to transition to test+trace.
By the way, we don’t have to be doing this badly. You can’t include a lot of graphs in a Facebook post, but go take a look at Italy, Spain, Germany, etc. in https://www.worldometers.info/coronavirus/and look at their death trends. They peaked and then went down already. The US and UK death trends are the only major countries that look roughly flat. (And the UK is run by Boris Johnson, who has been called the Trump of the UK, so this is likely not a coincidence.) You can also see that the US and UK have almost as many daily deaths as the rest of the world combined! We could and should be doing better.
How many deaths total will we have? My friend David Caulton has done a very nice analysis of the USA data, https://caulton.org/2020/05/03/70000-us-corona-virus-deaths-brace-yourself-for-some-sad-news/and he thinks that 142,000 deaths would be a reasonable (optimistic) estimate. His analysis is pretty simple, but I think good – he simply looks at the downward trend in USA deaths since the peak, 2.33% per day, and estimates out. Given that most of the country has been under quarantine for over 6 weeks, we’re very likely seeing the effect of quarantines on death rates right now and that seems like a great way to estimate to me. Unfortunately, if David had excluded New York, or factored in the relaxing of quarantines in a lot of places, he would be estimating massively higher numbers. (David calls his own estimate a lowball for this reason.) A lot more deaths are coming.
So, should the country be opening up now? The answer seems to me to be pretty clearly “no” at least not if our goal is to transition to test and trace. If you are making very slow progress, and you open up, you’ll move to either no progress, or even upward trends. You’ll never get to test+trace if you open up at this stage.
If some places like Washington are making good progress, but the country overall is flat, that means some other places are going the wrong way. The two most notable places are Texas and Georgia. Texas doesn’t have that many deaths – yet – about 224 (so per capita, very similar to Washington). But the difference is they’re headed up (about 30% per week) while we’re headed down. And the other difference is that Texas opened up recently, including e.g. restaurants. And not in a particularly smart or careful way (one large restaurant chain is actually forbidding workers from wearing masks, and firing them if they do.) So, if they were headed up 30% per week before, expect an acceleration. Georgia was headed up about 17% per week, and also opened up, including notoriously bowling alleys and tattoo parlors. These poor choices in Georgia and Texas impact us all – through the use of scarce treatments, tests, etc. and the potential for infecting other states as travel reopens.
I want to go back to discussing New York for a minute. They are rapidly headed down, but I doubt they are a success story. Estimates are that perhaps 15% of New Yorkers were infected with COVID-19. Some people, e.g. essential workers, and especially healthcare workers are at much higher risk than others; plus stupid people. That 15% of New Yorkers overall might represent e.g. 30% of those high risk categories or even more (just a guess). That’s enough for some sort of herd immunity effect – the spread among the high risk people to each other, and then to others they come in contact with, is likely down 30% or more. There’s also a “scared straight” effect – seeing people you know dead or in hospitals is going to make any quarantine more effective. But it comes at a terrible cost – lots of people dead, and lots of others infected (and possibly suffering long term effects.) It’s also NOT herd immunity. Once things open back up, there are plenty of uninfected “regular” people who can still be infected. The New York model isn’t one we want to apply widely, unless we want New York levels of death widely (around 330,000 deaths if we did this nationwide, and that STILL isn’t herd immunity.) Right now, Massachusetts and New Jersey are on track for New York levels of deaths, but at least they are, I hope, close to their peak. It’s unfortunate that other places like Georgia and Texas, which so far have been relatively safe, seem on track for this horrible outcome.