What the latest Covid blood donor survey shows
The South African National Blood Service (which manages blood supplies for eight provinces) and the Western Cape Blood Service have tested some donors for Covid antibodies for the past year or so. This has contributed to our understanding of the number of people infected with SARS-CoV-2 (the virus that causes Covid) and the proportion of infections leading to death. It can help us plan future waves, even if the exact way is complicated.
Assuming another wave towards the end of 2021 was almost inevitable – but before we had all heard of omicron – it was decided to carry out further such tests in early November. The numbers are now out.
The key results are:
- Overall, about 80% of black donors had already had Covid and 40% of white donors.
- There is no significant variation between age groups and genders.
- This latest survey did not include data from the Western Cape.
The test used does not detect the antibodies produced in response to vaccination, so it is indeed an estimate of the number of people infected.
Although blood donors are not perfectly representative of the population of the country, we can take into account the differences between the racial distribution of the donor population and the racial distribution of the general population. This means that our national estimate at face value is that around 70% of people had been infected before the omicron wave hit.
Since then we have had the omicron wave. We’d love to know how many people are infected now, but there’s really no easy way to derive that number. The researchers are currently updating their models with this additional data, and we may see estimates soon.
With that caution, here is my estimate on the back of the envelope:
- Omicron appears to have little trouble infecting people who have been infected with other variants, although there is some protection against previous infection and vaccination.
- By the end of last year, just over half the population had had a previous infection.
- Therefore, I estimate that about half of omicron wave infections were in previously uninfected individuals.
- Given estimates of the infection detection rate from previous waves and a number of plausible sources of possible variation in this rate, I estimate the detection rate to be around 1 in 10.
- Taking into account the approximately 700,000 cases declared between mid-November and mid-February, we obtain an estimate of 7 million cases, and therefore 3.5 million new infections.
- Given our population of about 60 million, that’s about an additional 6%.
Conclusion: it is not crazy to estimate that around three quarters of South Africans have now been infected. But I wouldn’t be surprised if serious models come up with even higher estimates.
A disturbing result of the survey is that once again it shows the serious racial disparities in South Africa. I don’t know if it affected the omicron wave. Estimating the racial distribution of infection after omicron depends intricately on variations in housing, lifestyle, access to vaccination, and all the usual factors that shape daily life in our country.