Coronavirus Outbreak – Total Tests and Positive Cases are Negatively Correlated

(Click on the image to enlarge)

Sample Extraction: 
1. Worldwide sample
2. Countries with total cases > 25K
3. Tests per million population > 10K

Given these criteria, many densely populated countries like Brazil, India, Iran, Mexico, etc. have been excluded. Since China does not disclose its testing credentials, it didn't make the list either.   


Highlights...

1. The correlation matrix (bottom graphic) shows that the total number of tests administered and the rate of positivity are negatively correlated, meaning more tests will not necessarily produce more positive cases. Of course, certain countries might buck this norm. For instance, though the USA has thus far performed the maximum number of tests, it has a high positive rate at 14%, defying the norm. Conversely, Russia, Germany and Italy -- the next three highest testing countries -- have positive rates that conform quite closely to the norm.

2. The total cases and total tests are highly correlated as only the symptomatic candidates -- a medically homogeneous population -- are being tested. This correlation will decline over time as the testing extends to some of the asymptomatic portions of the population as well. 

3. The total tests are also highly correlated with total deaths (0.7822) and active cases (0.8903), proving that the more serious or critical candidates have been receiving priority in testing. Considering the critical nature of those cases, they also remain active for a longer period of time.

4. As expected, the countries with the lowest actual death rates (total deaths to total cases) have also the lowest death rates per million population. For instance, the three countries with 1% death rates -- Saudi Arabia, Russia and Chile -- have the lowest population death rates at 7, 14 and 17, respectively.

5. Likewise, the countries with the highest actual death rates -- Belgium, France, Spain and the UK -- have the highest population death rates (> 400 per million) as well.

Stay well!

Data Source: https://www.worldometers.info/coronavirus/

-Sid Som
homequant@gmail.com