A tale of two or three cities (more different charts!)

The question of the week is if Georgia and other states should be reopening.  Naturally the response of most people breaks down over predictable political lines, but lets see if we can learn anything from looking at actual data.  The key question is if the available data shows that the spread is contained.  And that means looking at testing data.  As noted previously there are good reasons to be very suspicious about some of the testing in terms of timeliness and quality.  But let’s assume the testing data can be taken at face value and see what it tells us.  First, how widespread is testing?  Not very.  Here’s a plot of testing by day for Georgia, New York, Michigan, and Washington State.  Pretty depressing: they are all under 4% of the total population.  That’s insane considering that the SARS-COV-2 virus could have already spread to 5 to 40% of the country by some estimates …

Georgia is last, well under 1%.  But, if the virus isn’t widespread and is contained, that’s not inherently a problem.  If you’re testing lots of people, and those people are not getting sick or are negative, then you’ve expanded your testing outside the group of sick people and can be reasonable sure that you’ve got a handle on things.  If the positive to total ratio is under 10% I would start to be comfortable. A great sign is if that number is starting to trend sharply down over time (eg the more you test, the more negatives you find).   Let’s see what that looks like …

Oh.  That’s not great. Being flat, above 20%, means that GA, MI, and especially NY are not really expanding their testing outside the “known to be at risk” population enough. Washington State has been stable for a long  time in a reasonably good place.

Another metric we can look at is the ratio of tested positive to hospitalized.  Michigan and Washington have problems in that data, but Georgia surprisingly does have it …


What we can see here is that both NY and GA have been pretty constant over the last month in that around one in five people tested is either in or about to be hospitalized.   This is not what we would expect to see from a disease that is under control, or a testing plan that is expanding tests outside of people at risk and getting a handle on the extent of the disease.

So what does all this mean?  For Georgia it means the testing (which we are reasonably sure has issues with both application and interpretation) is inadequate to really assess what is going on.  The testing is just not widespread enough to tell.  Even in New York, where the number of cases per day is declining, and has tested 4 to 5 times as many per capita as Georgia, isn’t seeing that big drop in positives per tested population I would like to see before having some confidence in all this.  The governor of Georgia, being a good man of business, wants to get things restarted.  But that requires answering one key question.

That ultimate question, of course, is “how bad is COVID-19 anyway, compared to the draconian measures needed to stop the spread?” That really hasn’t changed much these last few weeks.  We know it spreads easily and rapidly.  We know most people don’t get very sick from it – but the people who do get sick tend to get really sick, require lots of medical attention, and a high percentage of them don’t make it compared to other respiratory illnesses.  That makes this especially hard to deal with.  If we had adequate testing, and aggressive contact tracing and isolation, we wouldn’t have to take generic intrusive measures.  Simple hygiene, basic precautions would suffice, with more precautions for the most vulnerable.  But the US waited too long to do these things, especially having a serious, widespread testing ability, so now we’re stuck.  Georgia is leading the way with an experiment to see just “how bad is it” with respect to the “surplus population” of vulnerable people. We’ll see in a couple weeks if the experiment pays off.  Unfortunately, by the time you realize the gamble hasn’t paid off, it’s too late.

It’s really hard to avoid Dickensian references when talking about the COVID19 pandemic in the US.  That probably means something.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.