Few nations have the resources to provide universal testing, which leaves governments with a difficult choice.
If they decide to test individuals with symptoms and the most unwell, treatments can be targeted to those who most in need. This has clear clinical significance.
But that approach can conflict with another important aim: to gather the best statistical information on the spread of the disease. To achieve that goal, countries ought to be testing random samples of their populations.
The reason to prefer random testing is that it provides an unbiased estimate of both the number of people that have the disease, and also the number of people that will die from the disease.
If countries test only those with symptoms, then the proportion of tests that come back positive increases, while the proportion of those tested who will also go on to die increases as well. There is no testing of individuals who have either mild or no symptoms but do in fact have the disease, and these individuals are never recorded as having survived or recovered from the virus.
This effect can clearly be seen when we compare the proportion of tests that come back positive to the proportion of patients that die from Covid-19 across different countries, as the following chart shows.
A clear positive relationship suggests countries that may have focused on testing the most unwell patients, such as Italy. Whereas those countries without a clear relationship, such as Australia and Germany, may have tested more broadly.
As this chart makes clear, the simple death rate being observed in each country around the world could be as much a result of the respective government’s testing policy as an actual property of the virus.