How Do You Like Your Numbers, Sir? (II)

But even when the numbers are correctly reported and tabulated, statistics can be tricky things.  Take an example from the metropolitan area where I’m staying.

The population of the greater metro area is just under 1,000,000.

The number of confirmed coronavirus cases in said metro area (total, since March 23) is just over 9,000.

How many of those people are sick now—or rather, how many are sick now, and don’t know it, and therefore aren’t isolating?

Unfortunately there’s a lag between when the data comes in and when it’s recorded in granular form, but the last week for which we have daily case counts is the week ending July 10.  If you assume that people are sick for about a week before developing symptoms, and assume that next week the reported cases will be the same as the week of July 4-10 (unlikely, since cases have been increasing; but one can always hope!), then that gives 385+221+420+773+575+795+662=3,831.  That’s how many presymptomatic, contagious cases there were (ish) in the last few days of June and the first few of July; and, since numbers are going up, that gives us a sense that a lowball number for presymptomatic, contagious cases right now would be about four thousand.

What about asymptomatic people, who never realize they were sick?  The CDC estimates this applies to 35% of cases.  WHO estimates 16% of those true asymptomatics can transmit the disease.  That means our supposed 4,000 current cases are actually only 65% of current cases.  If 4,000=65%, then 35%~2,154.  16% of 2154=345.  So there are about 4000+345=4,345 contagious cases in this metro area now, counting all the presymptomatics and the minority of asymptomatics who are likely to be contagious.

4,345 people out of 1,000,000 is 0.43%.

That’s actually quite a small number.

Another way of saying that is that if you walked around this metro area bumping into members of the populace at random, 99.57% of them would not have the virus.

Not bad, no?