If testing is limited but effectively randomly sampling, then measured incidence will be proportionate to actual incidence.
If testing is targeted toward regions of higher interest and likelihood, then test positivity should skyrocket out of proportion to actual positives. (I'm not sure what the case is for South Africa.)
Using deaths as a lagging indicator of cases, South Africa reports about a 3% CFR (based on reported cases and deaths), as opposed to about 2% for the US. This would suggest a somewhat lower testing prevalence in ZA, by about a third, but not an especially bad record. This does assume that Covid deaths are being accurately assessed and reported. Total excess mortality is the usual check for that.
My read is that ZA's testing infrastructure is reasonably good, and that the B.1.1.539 variant's growth is extraordinary.
If testing is targeted toward regions of higher interest and likelihood, then test positivity should skyrocket out of proportion to actual positives. (I'm not sure what the case is for South Africa.)
Using deaths as a lagging indicator of cases, South Africa reports about a 3% CFR (based on reported cases and deaths), as opposed to about 2% for the US. This would suggest a somewhat lower testing prevalence in ZA, by about a third, but not an especially bad record. This does assume that Covid deaths are being accurately assessed and reported. Total excess mortality is the usual check for that.
My read is that ZA's testing infrastructure is reasonably good, and that the B.1.1.539 variant's growth is extraordinary.
I'm relying on Worldometers data:
https://www.worldometers.info/coronavirus/country/south-afri...
https://www.worldometers.info/coronavirus/country/us/