There's a lot of graphs and stats that the news is throwing at people right now. So much so, that you can get information overload trying to make sense of the statistics that have meaning. To quote my old Econometrics professor, "There are three types of lies: 'Lies', 'Damned Lies', and 'Statistics' ". I should also lead with the caveat that I'm an engineer and data nerd by trade, but I'm not an epidemiologist. I welcome feedback from those who have more experience than I do.
The most important question we're trying to answer (at least here in Michigan), is "How are we doing?", and "When can we reopen our economy?". With respect to those questions, here's my take on the most important data, and some caveats about what these data are telling us.
The four most cited data in news stories are:
- Total Number of Cases
- Daily New Cases.
- Total Number of Deaths
- Daily New Deaths
This post will talk about #1 and #2 above. I'll get into #3 and #4, and why it's important for us to understand the real death rate of infected cases in an upcoming post.
The short version? Total numbers don't tell us much, but sound sensational on the news. Daily numbers tell us how we're actually doing in our fight against COVID-19
#1 Total Number of Cases
From a decision-making standpoint, this number is prety much meaningless. It's a score-card of sorts, but higher population areas will have higher number of cases, even if the rate of spread is better controlled in those areas. I'm breaking down most areas by State in this analysis, because it's the easiest numbers to get, and because most states have a uniform set of rules for all of their areas at this time. (Some states are opening by district or county, and some cities have different rules than the state at large, but I won't be going into that too deeply.)
So, when you see one of those commercials that has a spooky narrarator talking about how the USA has "the most coronavirus cases in the world", know that we have one of the largest populations in the world. It's to be expected. It's not spreading in the USA any worse than it was in, say, Italy or France. The standout exception has been China, which has many more people than the USA, but had fewer cases of COVID-19. It also has a draconian government that can ruthlessly lockdown their population. America, not so much (thankfully.)
#2 Daily New Cases
This is the data that is most relevant to our policy decisions. Daily New Cases will tell us how much COVID-19 is spreading in our states. In fact, even more important than the number of new cases is the trend line. Is the graph rising or falling? If the graph for Daily New Cases is trending upward, that means that the spread of the virus in a given state is still growing exponentially (If the new cases today are higher than the new cases yesterday, it's spreading faster than it was yesterday). There's an excellent youtube video on exponential growth from 3Blue1Brown I encourage you to watch here: https://youtu.be/Kas0tIxDvrg.
The most important thing to note is: it doesn't matter how many total cases you have in a given state, it matters how fast it's spreading. If your total cases are low, but your daily new cases are trending upwards, you're not doing better than other states, you just started later than they did. Get ready.
I'm going to compare two different states with similar Total Cases, and slightly different approaches on stay-home orders. I'll look at Michigan's stats first, because it's home, and the state I've been watching most closely. Michigan, compared to some other states, has had fairly severe stay-at-home orders in place. Here is the daily new cases, and you can see that, after the initial spike in new cases, the trend line is moving gradually down on average. This means that tomorrow's new cases will be fewer than today's. Michigan, it seems is heading in the right direction.
|Michigan NEW DAILY cases|
Now let's take a look at Michigan's TOTAL cases, you can see the tell tale Logistics curve, that starts exponentially, but then inflects and starts to flatten out. This is what "flattening the curve" means.
|Michigan TOTAL cases|
These two graphs are directly related, As the top graph approaches Zero new cases, the bottom graph will totally flatten out to horizontal and will stop growing entirely. At that point, Michigan has no more new COVID-19 cases. Here's hoping.
So if we look at the State of Texas, for example, which has had (relative to Michigan) somewhat more relaxed stay-at-home orders for most of the state. You can see fairly shallow steady linear trend upwards. This means that, for the state as a whole, there is exponential spread in the state still present. Not only are the number of cases growing, but they are growing a little faster every day. Think of it in terms of compound interest, and what happens to your credit card bill when you only pay the minimum balance.
|Texas DAILY NEW cases|
Texas's curve, by comparison to Michigan, is not yet flattening...
|Texas TOTAL cases|
Michigan and Texas have similar total number of cases (as of the writing of this post), even though Texas has almost triple the number of inhabitants. But if the trend lines continue, Texas will easily surpass the number of COVID-19 cases in Michigan. On the other hand, if Texas bends their curve similar to Michigan, their total cases will end up similar to ours, meaning their per capita cases will end up much lower than Michigan.
Lag Time between Policy and Effect
Take a look at this graph from Michigan again. It's worth noting the time between when Michigan's Stay-At-Home orders went into effect (1), and when the curve reversed direction (2). Note that it took about 3 weeks for the graph to reflect the stay-at-home orders. This will be important as we make changes in policy and open up our economy.
|Lag Time between behavior change and visible results|
We won't know how effective we are for 2-3 weeks weeks after a change. So as of 5/26/2020 retail has opened up, and in-person gatherings of fewer than 10 have been allowed. Also, some larger churches, and notably, the Archdiocese of Detroit, have opened up church services in some limited capacity. So somewhere around the end of the current stay-at-home order rules, June 12th, we should have reliable data on how well these changes have been received, and whether or not we're still trending in the right direction. If the number of new cases starts to spike upwards again, expect to see some restrictions come back into effect.
Post a Comment