How the Coronavirus developed in China and the implications for the US

This chart, which shows how the coronavirus developed in China, may be the best, most important one I’ve ever done.

 As you know, the virus follows a fairly predictable pattern that can last for more than a month. First, a certain number of people get infected. Then, over many days, some of them begin to show symptoms. Then, more days later, some need to be hospitalized. Finally, often weeks later, patients either recover (albeit some with long-term issues like lung damage) or die. There’s a lag effect – like a pig in a python. It’s critical to understand this – but most people don’t.

 In this chart, for each person who tested positive, there are three dates: when they said they first started feeling sick (the blue line shows the total number of such people in China each day), when they tested positive (red line), and when a small fraction of them died (green line). (Each data point is a three-day average.)

Source: JAMA (symptoms) (original source here); Worldometers (tests and deaths)

 For each metric, you can see a surge and then rapid decline, but at different times.

 The number of new people showing symptoms peaked on January 26, but the peak in the number of new cases per day wasn’t until 11 days later on February 6.

 This reflects two things: first, most people don’t rush to get tested the day they first start having symptoms like a mild sore throat, cough, and fever. Most probably figure it’s just a regular cold, so they don’t seek treatment – much less get tested – until they’re really feeling lousy a week or two later. Then, compounding the delay, there were testing delays in China, just like we’re seeing here.

 And the peak in deaths wasn’t for another eight days on February 14.

 Look at what happened in China (mostly the city of Wuhan): the number of infected people – the gray bars – rose rapidly, about 10 times, from 190 to 2,008 (trailing three-day average) – in only nine days from January 14 to January 23. That’s when the Chinese government did an about-face and implemented an extreme lockdown of the entire province, shown by the purple vertical line in the chart above.

 And it worked – almost immediately! As you can see, the number of newly infected people only rose slightly the next three days, peaking at 2,717, and then, in less than three weeks, steadily dropped to almost zero!

 But here’s the key: nobody could see this.

 Instead, people could only see the number of reported new cases – the red line. Looking at this data, one would no doubt be convinced (wrongly!) that the lockdown wasn’t working because reported new cases were skyrocketing.

 To repeat: whereas the actual number of new cases peaked almost immediately after the lockdown, the peak in the reported number of new cases didn’t occur for another 14 days!

 In summary, as soon as China locked down Hubei province, the growth of actual new cases plateaued and started to fall within days, dropping to almost zero within three weeks. Yet reported new cases soared for another two weeks as testing caught up with what was really happening.

 And deaths – which is what really generate headlines and panic – didn’t peak until eight days after that.

 I’m cautiously optimistic that the same phenomenon is playing out right now in the U.S.

 I can’t prove it, though. It’s an educated guess, nothing more.

 The reality is, until there is more widespread testing and we get another week or two of data, we’re all flying blind. Nobody knows how many people are actually infected here in the U.S. right now, where they are, and how rapidly that number is growing.

 Because we can’t answer these questions, those who are making the most extreme predictions are getting the most attention – and nobody can prove them wrong.

 But what I do know is, because of the lag effect, even if the measures we’ve taken in the past couple weeks are working well (as I believe they are), bringing the virus’s replication rate well below 1.0, we won’t know it.

 As a result, the number of new cases, hospitalizations, and deaths will almost certainly continue to go up for at least a couple of weeks, almost certainly leading to universally bad headlines in the U.S. and Europe during this period.

 So, fasten your seatbelt – it’s going to be a wild ride.

 But don’t panic and don’t lose hope!