Is 1 million tests good?

The US hit a major milestone in terms of testing yesterday, as per the Covid Tracking Project over 1 million covid tests have been conducted in the country. This is by far the most of any country that reports these numbers, and considering how late the testing started, it is truly an impressive feat. However, does this mean that we have a good idea of the spread of the virus in the US? There are a few statistics I track that might give us a better indication than just a pure top line number.

Country# of testsTests per 1M in pop% positive
Norway90,242166455.14%
Italy477,359789622.16%
S. Korea395,19477132.18%
Canada244,42164713.52%
Austria49,455549019.79%
Denmark27,109468210.55%
USA1,048,971317117.97%
UK152,979225116.44%
# of tests completed as of March 31, 2020 – various official sources

The first obvious thing to do is to scale the number of tests by the size of the population. If I’m in Luxembourg 1 million tests would mean I’ve tested my entire population, twice, but if I am in China that’s not a significant percentage of the population. The US has so far tested approx. 3,200 people per 1M in population, which is the same as saying it’s tested 0.3% of its population. Given that they only started testing a couple of weeks ago, which is far later than all the other countries in the list above, that’s a pretty impressive pace, but in the end it still falls short of where it should be. There is a lot more work to do to get a good sense of where the pandemic is.

Canada is testing at about the same pace as the US per 1M in population right now (roughly increasing by 500 ppl per 1M per day), but we’ve had a solid few weeks in terms of a head start, so we are in a much better shape overall at 6,500 people tested per 1M in population.

The other key statistic to track is the % of positives identified across all cases tested. In general the countries that are testing sufficiently and are able to contain the outbreaks have lower positive rates. This makes sense because they are testing more people who have symptoms but are not covid positive, which in turn means they are not just testing the sickest of the sick but have the capacity to test beyond that. Additionally if the sample can be assumed to be somewhat random, the positive % should be relatively lower for countries where the virus is not as wide spread.

We can see that Italy has a very high positive % despite having tested a large portion of its population. This is not good because it implies they either don’t have the testing capacity to reach the symptomatic people that would test negative, or the country is in fact that infected. The situation there is improving daily, and the rate is starting to come down slightly, so I expect we’ll know the answer soon.

On the other hand the US positive % rate is almost as high as Italys, and it is still increasing daily, which are both very troubling signs. However, as bad as the overall US rate is, the NY state rate is close to 40%!! That means that for every 2 people that get tested in NY state, 1 of them tests positive for covid. That’s truly frightening, and I believe the other major cities in the US are not yet testing enough, so unfortunately I think the worst is still to come.

The situation in Canada is not all roses, as the positive test rate is still climbing, but at 3.52% overall and 6% in worst hit Quebec, I think it’s relatively speaking under control. We will know more when the backlog of test cases is cleared, which is on schedule for the end of this week or latest early next week. I’m cautiously optimistic, but a jump in the positive rate would be an early worrying sign.

Photo by CDC on Unsplash


Follow me on:
Facebooktwitterrss

1 thought on “Is 1 million tests good?

Let me know what you think