3 days ago I posted a projection of the increase in cases of coronavirus in the US based on a 3 day doubling period, which was about what we were seeing at the time. We had about 2500 cases then. This morning the tracker I use (more about that later) shows us at 5286 cases. For now, we’re still on track to have more than 160,000 cases by April 1.
Here’s the takeaway. It took us nearly two months to go from the first US case to 2500 cases. We went from 2500 to 5000 in three days. Our next 5000 cases will also take about three days if this progression continues.. This is the nature of exponential curves. Each day adds a percentage of the previous day’s totals, over time those percentages compound. When this is all over, there’s an important lesson about compound interest in here.
So what’s the difference between 5000 and 5286? How much difference does that make in the math? Here’s a comparison of a projection to April 1 based on a 3 day doubling starting from 5000 and 5286:
5000, 10000, 20000, 40000, 80000, 160000.
5286, 10572, 21144, 42288, 84576, 169152
An increase of 286 in today’s cases leads to over 9000 more cases on April 1. Every case we can avoid today, reduces the projection by 32 cases on April 1. If we avoid 3 cases today, we reduce the April 1 number by 1000. That reduces hospitalizations by 20, and saves 2 lives. That’s why it’s so important to act as early and as effectively as possible. That’s why all of the bans and restrictions are necessary.
One thing to remember is that the effect of all these measures won’t show up immediately. First, reported cases were probably transmitted before these measures went into effect. It takes time for symptoms to show and tests to be run, so there’s always a lag in seeing the effects. Second, as testing becomes more available, we’ll see increasing numbers just because more actual cases are detected. Even if we’re really slowing the actual transmission rate, the reported rate may stay the same or rise. This happened in China’s data when they changed their criteria for testing.
Technical details and such after this.
These numbers are just reported cases, not actual cases. Nobody knows how many actual cases there are, but it’s certain that the number is much, much higher. I’m using reported cases because it’s the only data we have. Testing is not widely available yet and many cases are mild enough to not involve any kind of treatment or assessment.
There are many trackers and they all show somewhat different numbers. There is no single reporting system, and trackers update differently. The WHO tracker shows the US with only 3536 cases, for example. For this purpose, it doesn’t really matter which tracker we use, but it does matter that we consistently use the same one. I use https://coronavirus.1point3acres.com/en . I like its display options, it updates very frequently and is very prompt at counting reported cases. I especially like that it documents the source for each report and actively curates the reports for duplications and accuracy. When I have double checked numbers, it has turned out to be accurate. As you see different numbers, remember that reports take time to work their way through the systems.
Finally, this is not meant to be an academic study. I’m using round numbers to make a point about the importance of action and to prepare minds for what may lay ahead of us. One way to avoid panic is to have an idea of what may lie just around the bend. The exact numbers will be somewhat different, though should be in the ballpark.
I’m happy to address questions or concerns in the comments. Feel free to share if you like.
Stay healthy, physically and mentally.