Covid-19: Excess deaths for second week running

The data for each week shows the lower and upper prediction for the number of deaths and the actual forecast number which is halfway between the lower and upper bounds. This is calculated based on historical data.

Excess deaths, which refer to the difference between the observed number of deaths and the forecast number of deaths based on historical patterns, have exceeded the upper prediction for two weeks in a row.

Last week the MRC reported that the number of weekly deaths exceeded the upper bound of historical forecasts after a few quiet weeks where the number of deaths were only slightly above estimates based on historical data, but below the upper bound.

In the week from 14 to 20 October, the MRC forecasted 8,883 deaths. However, it recorded a total of 10,225 deaths, an excess of more than 1,300 deaths.

The number of natural deaths (i.e. not homicide, suicide, or accident) for the week ending 20 October was 9,080 deaths, with an excess of 1,254 deaths.

The number of unnatural deaths showed a slight decrease in the last week, and has remained within the prediction bounds.

The overall number of excess deaths from all and natural causes both increased for the period between 6 May to 20 October. The total number of excess natural deaths now stands at 47,923, while the number of excess deaths from all causes is 41,246. (The all-cause deaths are lower because unnatural deaths have been lower than historical forecasts.)

Far below the peak

All provinces recorded deaths above the weekly forecast. The number of excess natural deaths are still below the upper bound though, except for the Free State, Mpumalanga and the North West.

The excess deaths are however very far below the peak of the South African Covid epidemic in July. The rate at which the deaths have increased is also much lower than July, suggesting that the current surge in deaths is much less serious than then.

It cannot definitively be said that the excess natural deaths are all Covid, but it is likely that the vast majority are. It is probable that the number of Covid deaths in South Africa since March has been between 41,000 and 48,000. The number of confirmed Covid deaths is slightly less than half this at 19,111. Nearly all countries undercount their Covid deaths because many people die without having been diagnosed with the disease.