Omicron in South Africa

Illustration of south africa flag Free Vector

Omicron (SARS-CoV-2, B.1.1.529) was identified on November 24, 2021 in South Africa (SA). It’s been somewhat downplayed as being “mild,” although it’s too early to make that claim. It’s already clear that it is quickly transmitted and that the COVID vaccines may not be as effective. Some data is available now to get a better sense of what this variant could mean for the rest of the world.

First, it’s worth looking at the scope of previous waves of COVID in SA. The scale on the left side of this graph is for the blue portion of the graph, which is known as the epidemic curve. It also has a dark blue line representing the seven-day moving average, which is the way that many media outlets display the data.

The gray part of the graph is the deaths in the same format, but the values for that are on the right axis.

These are cases and deaths for ALL variants of COVID in SA. The most important characteristic in this graph though is to note the somewhat gradual way prior waves climbed to their peaks. The difference with the current wave at the far right is that it is almost to the same height in only a couple of weeks, whereas the other waves took almost two months to reach that same height.

The other important part to note is that generally speaking, increases in cases don’t start showing up as increases in deaths until 2-3 weeks later. In the second wave, that doesn’t appear to be the case, but that is likely due to this being the entire country in one graph and a number of reasons that could explain this pattern.

The next graph is the same blue epidemic curve, but now contains something from calculus known as the first derivative. This particular derivative uses the past three weeks of data to calculate the rate of growth or slowing in cases and is the red line in this graph.

The best analogy to the red line is to think of it as the accelerator or brake on a car. There is a black dotted line running across the graph at zero on the right axis. When the red line is near this line, there is relatively little change in the number of cases at a given time. When the red line goes above the black dotted line, cases rates are accelerating, and the higher the red line is above the dotted line, the faster that acceleration is happening, similar to flooring the accelerator.

The opposite is also true. The lower the red line is below the dotted line, the faster cases are slowing, similar to stomping on the brake when the red line is at the very lowest.

One benefit to this approach is being able to project a short time into the future using this derivative. Just like with a car, the faster it is accelerating, the longer it takes to slow down. I’ve found that this three-week derivative can project about 10 days into the future with reasonable accuracy, which is represented by the yellow highlight on the far right side.

If you look at where the red line is vertically above where the blue epidemic curve ends and compare it to the peak of the red line in the other waves, you can see that with current data, the rate of new cases is already at the highest it’s ever been in SA. That rate is likely to keep climbing over at least the next week.

The red line in the next graph is the exact same derivative as in the previous one. The black line is what is known as the second derivative. The simplest way to explain it is a measure of how fast the change of the rate is changing. This can be used to identify if the first derivative (red) will continue on its current trajectory, climb faster, or start to slow. If you look at the yellow area again, you can see that the second derivative remains close to zero currently, which indicates that the rate of growth of cases isn’t changing anytime soon.

The next graph has the same blue epidemic curve, but now has a number of colored lines representing each province in the state, again measured on the right side of the graph. These show the number of hospital admissions in each province.

The province of importance with omicron is Gauteng (the red line). This is the one that contains Johannesburg, Pretoria, and Soweto. Again, at the far right, look how quickly hospitalizations are climbing in the province. that provides a great deal of evidence that omicron is NOT mild. It’s also apparent that some of the other provinces are starting this climb as well.

Further evidence that this increase is due to omicron having become the dominant strain in the country. Omicron had not been detected in any sequenced samples for the period ending 11/1/21. By 11/15, it was 25% of samples, 11/29 it was 85%, and for the period that will end 12/13, it’s currently 100%. This is well visualized at, which is the source for the following graph. It shows the proportion of any particular covariant over time. Omicron is the purple area at the far right.

The final graph that supports this interpretation has the same background epidemic curve. The right axis on this graph is hospital admissions and discharges. The green line is admissions, the red is discharges. Again, there is about 2-3 weeks between these curves. The reason that the numbers are different between the curves is the number of deaths.

Again, looking at the far right of the graph, it’s far too early to detect deaths. Most likely, the earlier decedents did not have their samples sequences, so it is reasonable to assume that there are deaths from omicron, but they simply have not been detected yet.

So what is the takeaway? This could easily be far worse than delta, which we are really just starting to feel the impact of in the US and will really start to climb around Christmas. The best thing to do to understand this variant is to continue to monitor the impact in SA but also to watch the United Kingdom closely, because that is likely to be the next area most greatly impacted by the variant.

As an individual, get your booster, wear a mask (preferably a respirator), and don’t spend time indoors with anyone outside of your household. Of course it’s the holidays, but is it worth making it the last one where you might see loved ones alive? This is incredibly serious and we are failing in our response as a country. Do your part for the good of those around you.

12/19/21 Here’s an excellent additional Twitter thread on omicron showing why it will likely be worse that what we have seen from other variants in the US.

Leave a Reply