SA, the “tests per million population” might be misleading. Back in last spring I read of “pool testing” in an African country – this was a blood test, I think. Everyone in a community group gives a blood sample. A portion of each sample is taken, and the portions are then combined to form the “pool”. This pool is then tested. If it’s negative, no one in the community group is infected. If it’s positive, there are still the remainders of each sample that were not combined, for narrowing down who’s infected and who isn’t.
If the WHO have been coordinating, this technique might be widespread by now. It is particularly well suited to nucleated communities.
Does the WHO publish assessments of testing coverage?