Yesterday I pointed out that South Korea appears to be finding more of the mild or asymptomatic cases that other countries are missing. But they've tested at most 0.6% of their total population. Their official website says you are eligible for testing if you have fever and/or respiratory symptoms within 14 days of being in contact with a known case. So if you are totally asymptomatic in South Korea, you're not being tested.
But there are a few instances of testing a whole population that give us some interesting data.
First, is the town of Vò in the region of Veneto, Italy. Vò was one of two initial outbreak locations in northern Italy. Upon finding their first two "obvious" cases in late February, including one death, the town and surrounding area was tested in its entirety.
Unfortunately the numbers being reported are not entirely consistent, but the best I can reconstruct is this:
On 2/28, it was reported that there were 133 positive tests out of 6800. Presumably this is the region of Veneto, not the town of Vò, because there are only about 3000 people in Vò.
That's a ratio of 67 hidden cases for each known one. There's obviously a lot of room for error bars. If it just so happened that three people had obvious cases, it would be 44, and if they had only one, it would be 133. It's not great statistically, but it's a rough estimate. This is at the very beginning of the outbreak. Of the 133, 100 were described as asymptomatic. Presumably some of those were actually in the incubation period and developed symptoms over time, but I haven't seen any clarification on that.
In the past few days, more news out of Vò has been released. Out of 3000 people tested in Vò, 66 were positive. That's a ratio of 33 hidden cases for each of the two positive cases in Vò. Those 66 people were quarantined, the town was blockaded from the world, and is now free of the virus.
Another relevant case is that of the cruise ship Diamond Princess, which was quarantined in early February off the coast of Japan, with 3711 on board. As people needed hospitalization, they were removed from the ship, and in early March, all the passengers were removed and quarantined on land. By Feb 20, there were 634 confirmed cases; 328 (52%) were asymptomatic. That's out of 3063 tested. Surprisingly, the asymptomatic cases skew older, although that might be an artifact of who was tested. Older people were tested first, and if you look at just those in their 60's and 70's, they're 58% asymptomatic.
As of March 15, the cases have increased to 712. I'm not seeing any data on whether any of the the previous asymptomatic cases eventually developed symptoms, or were truly asymptomatic for the duration.
To date 7 passengers have died. The case fatality rate on the ship was 2.3%, close to the 2.5% I estimated yesterday, although the average age was 58, higher than that of most countries. This suggests that my estimate is a little bit high.
Putting together what we know about Vò and the Diamond Princess, it appears that most of the hidden cases in Vò were not totally asymptomatic, which is good news. The bad news is that more than half of people who are infected at any given time are asymptomatic.
Because the US is still not testing very much, we are more like Vò than the cruise ship, with the actual number infected at 30-60 times the official numbers. That means the US currently has roughly 400-800K infected, with a bit over half asymptomatic, the official 13.8K in bad enough shape to have attracted attention, and the remaining few hundred thousand experiencing mild symptoms.
Assuming a 2% CFR, 8-16K currently infected people will die. Of course, transmission is still happening, so all these numbers are expected to go up.
But remember the town of Vò, which defeated the virus in a few weeks by testing everyone and quarantining the positives, regardless of symptoms. If we could do that, the virus would be stopped in its tracks.
Source: https://www.medrxiv.org/content/10.1101/2020.03.05.20031773v2.full.pdf
