SA’s lockdown ‘saving’ on overall deaths – Dave de Klerk

SA’s consistently low ratio of positive coronavirus cases to tests might mean our recently increased testing teams are struggling to find positive cases – a very promising development, if true. The caveat is the unproven efficacy and “hot-spot” coverage of our testing teams. This speculatively hopeful analysis is based on existing comparative data which the author mines in a unique way. There’s an even brighter light on the horizon. It centres on expert actuarial thinking that highly sought-after and still-to-come, random anti-body testing in our townships might show high exposure to Covid-19, but low infection. That’s based on the truth that many “poorer” African and other countries, show, (like us), very low coronavirus death rates. Actual deaths per million of each country’s population is supposedly the most meaningful statistic. Could childhood exposure to illnesses and hardships build up some natural immunity to diseases like this? For now, we’re grateful for our globally-stellar relative numbers (see Table). Here’s another gratitude; South Africa’s road accident and homicide deaths are usually way higher than most countries. Lockdown and the unavailability of alcohol has sent these plummeting – the “saving” may well exceed the relatively few deaths caused by the virus (123 to date). – Chris Bateman 

By Dave de Klerk*

Again during the past week the absolute number of positive cases in South Africa increased considerably but again, because of our significant expanded testing, the ratio of positive cases to tests barely moved from the previous very low figure. As can be seen from the table below the ratio on 2nd May of total positive case to date (6,336) to the total number of tests done (230,686) is still only 2.7%. To get a better idea of the ratio for newer tests, I’ve included a column of positive cases to tests done over the previous week (25th April to 2nd May). This shows a marginally higher ratio for South Africa of 2.8%. Our testers are still really struggling to find positive cases.

*All data obtained from www.worldometers.info. Figures for first 3 columns of the table for China, France, Spain, Netherlands, Sweden Brazil and Mexico havebeen excluded because their daily number of tests are not updated regularly, if at all. The fourth column showing deaths per million population for them is as follows. China 3, France 379, Spain 537, Netherlands 291, Sweden 264, Brazil 32 and Mexico 16.

The Western Cape’s ratio however remains significantly higher than the rest of the country, at about 7%.

For most other countries the table shows that their most recent week’s testing produced a lower percentage of positives than their cumulative figures show. In some cases (As in the UK) this is partly the result of much more extensive (voluntary) testing being done, eg offering their new drive-through testing facilities. Their week on week tests are up a massive 170% as can be seen from the table. In many other countries the lower recent test ratios probably do indicate that they are significantly reducing their population infection rates-presumably by their lockdowns. Countries showing the opposite trend-Singapore, Chile and Russia –possibly have a worsening infection rate, although Singapore’s reported testing numbers have reduced, which looks suspicious.

From the figures collected by Worldometers.info it appears that some countries report neither their testing numbers nor recovery numbers very accurately, or not at all (eg the UK doesn’t report recoveries at all). Deaths and positive cases diagnosed seem to be more accurately reported although there have been suggestions that deaths may often be under reported.

At the end of the day however, when the virus has run its course, the actual deaths per million of each country’s population will probably be the most meaningful statistic.

The current levels of that figure are also shown in the table as well as the ratio of cumulative deaths to cumulative positive cases found.

Once again South Africa looks good on both these measures.

Deaths to total positive cases 1.9% and deaths per million population only 2%.

Many countries project their expected average deaths month by month in a “normal year”

By comparing actual deaths this year with the expected numbers, an alternative estimate of the Covid-19 caused deaths can be obtained.

South Africa is in an interesting position in this regard. Normally road accident and homicide deaths here are much higher than in most countries. Because of the lockdown and unavailability of alcohol, these deaths are very low this year and the “saving” will probably far exceed the relatively few deaths caused by the virus (Only 123 to date).

Looking at the latest global recovery percentage (1,121,608 recoveries out of  3,484,502 positive cases, gives a percentage of 32.2). South Africa has done better than this with 2,549 recoveries out of 6,336 positive cases, or 40.2% – but as mentioned earlier recoveries are under-reported by some countries so the accurate global percentage may well be higher.

Another interesting finding relates to the age relationship of Covid-19.

When Chinese data was produced some time ago, I compared their age-related data of the virus with a mortality curve for “normal” mortality. The age relationship is very similar.

For example the Chinese figure for people over 80 was 15% and, depending on the age distribution of people over 80, a reasonable assumption also gives an average mortality rate of about 15%. Put differently, about 150 out of every 1000 people in this age range, will die each year from “normal mortality”. If they catch Covid-19 a similar number will die. What is the overlap is the 64 000 dollar question. It might be very high. From anecdotal evidence co-morbidity seems to play a very big role in Covid-19 deaths. The same co-morbidity doubtless would play a big role in determining “which 150” of the 1,000 are more likely to die in a “normal year”.

If one looks at the two columns in the table relating to mortality it does look as if some of the “poorer” countries are getting off relatively lightly and vice versa-so far at least! There are many more such “poorer” African and other countries not included in this table, which also show very low death rates. Could childhood exposure to many illnesses and hardships perhaps build up some natural immunity to a disease like this one? One way to test whether a significant number in our township people might have been exposed to the virus, but not become ill, would be random antibody testing and hopefully this will start being done here sooner rather than later.

• Dave de Klerk is a retired actuary living in, or rather locked in, Cape Town.