The COVID-19 Dance Phase : South African Contextualization

This article puts into context ideas put forward by Tomas Pueyo in an article titled “Coronavirus: The Hammer and the Dance”. The aim is to put a spotlight on The Dance Phase using a Daily Growth Rate approach instead of R-naught with a special focus on South Africa.

Source Disclosure: Graphs used here are my own manipulation and visualization of COVID-19 data obtained from Worldometer

Total Confirmed Cases in South Africa

Suppression : The Policy of Choice

According to Pueyo, there are two policy choices political leaders have in handling viral infectious diseases: Mitigation and Suppression. South Africa, like many other nations , apparently chose Suppression. Suppression involves two sequential processes. When the outbreak starts, authorities pounce on the virus, apply tough measures like social distancing and lockdowns to get the virus under control. This phase is called The Hammer. It is essentially “hammering” down the virality of the virus . The sole objective is to lower the infection rate so that the virus spreads around the country at a slower pace. Its an attempt to get things under control. If successful (it is in most cases), the infection rate goes down.

Once the infection rate goes down, the next phase called The Dance kicks in. The dance is essentially a long period of time where the authorities “dance around with the infection rate” until a vaccine is found. Total lockdowns and social distancing measures are very expensive. The dance is a balancing act. Balancing economic livelihoods and lives that can be lost to the virus. Authorities relax regulations when the rate goes down and tighten regulations when the rate goes up.The infection rate will be dancing around a certain level.

Infection Rate : Choosing to use Daily Growth Rate instead of Rate R-Naught

So obviously , the suppression is based on the ability to actively monitor and measure the infection rate. Epidemiologist and virologists use R-Naught (R0) which is a basic reproduction number of the virus. It is the number of people that can be directly infected by one infected person. For COVID, they say its around 2.4 and they factor this into models that predict how the virus will flow through society.

I am not a virologist. I am an accountant. I have a slightly different understanding of numbers and growth. I am familiar with interest rates, compounding , capitalization of interest and so on. Instead of using RO, I will use the Daily Growth Rate (DGR) as the key metric to actively monitor the progression of the virus. The Daily Growth Rate is basically the percentage change in the total number of cases measured on a day to day basis. If we had 100 COVID cases yesterday and 110 today, it means the virus increased by 10% in a day. We then focus our attention on this metric.

How is the Daily Growth Rate Performing?

Great. It has been decreasing for a while now for most countries though the absolute number of cases are increasing. Think of a decrease in the DGR as a decrease in the interest rate that the bank charges you on a million-dollar loan. Its good news when rate decreases, but you still owe the bank a million and the interest is still being added to your total debt.

Lets go through the DGR for the world and DGRs for selected countries with more than 10,000 cases. The hand-drawn downward-sloping red line represents the effect of the hammer. The hand-drawn flat horizontal line represents the dance phase.

World DGR, Decreasing everyday. Lockdowns around the world reduced DGRs from 8–10% region to 3–4% region

The DGR for the world has been decreasing but might increase as high population countries such as Brazil, Pakistan and India enter the rapid growth phase.

Next. USA Daily Growth Rate. It has also been decreasing since the beginning of the month.

USA DGR decreasing from 13% to 5% in April and now dancing around 4.75% due to lockdowns in NY and NJ and social distancing measures in other hotspot areas.

The USA rate could dance at an even lower level but the cost to the world’s biggest economy would be hard to swallow for Americans and Trump. It looks like it will dance around 5%. Trump is pushing for reopening the entire economy by end of May but it is highly-unlikely that he will actually do it. Doing so will shoot the DGR up to unmanageable levels.

Turkey below, looks like it will also converge to a DGR around 5%.

Turkey DGR , big decrease in April from 15% to 7%. It still decreasing but a slower rate. It could possibly dance around 5%

The Spanish DGR dropped from the 7% region to the 3% region and it looks like it will be performing a jumpy up-down dance for the rest of the month.

Spain DGR

The Italian DGR, which is ahead of most countries by a week or so, has been dancing below 3% since the 6th of April.

Italy DGR, leading by a week or so.

In all the charts above, the trend is that of a decrease. The UK DGR (below) seemed to be jumping up and down though it has lately shown a general trend to converge around 5%.

This could be a result of policy-confusion and policy flip-flops. You cannot clearly draw a downward “Hammer line” and you cannot confirm that the DGR has converged to a certain level with a high degree of certainty.

Compare and contrast the UK DGR with the German DGR (below) which is bumpy but still shows a discernible “Hammer Line” and a “Dance line”.

Germany DGR, dropped from around 8% to around 2.6%

Over a period of time, we should be able to observe more pronounced dance lines. By the end of the month (April) some countries should have relaxed their measures and we should see the DGRs rising, then in May we should see some measures being put back in place again to bring down the DGRs. Countries will learn to establish DGRs that are consistent with their localized conditions. For example German can afford a higher DGR because she has a better healthcare system with double UK’s ICU bed capacity.

South Africa’s DGR drop is phenomenal but….

South Africa GDR dropped from 26% to 7% and appears to be dancing within a wide band from 1% to 10%

The DGR for South Africa dropped massively because the a very big hammer was used to crush the virus. It was very aggressive and was done at a very early stage such that we have sporadic daily increases from a low base.

Presenting the same data from the 2nd of April to the 18th gives us a picture of the dance during lockdown.

South Africa GDR, dancing all over the place during the lockdown

But….South Africa is still at the early stages of the evolution of the virus. A better pattern should emerge once the number of infections reach 10,000. The current number of cases is just above 3,000 . The data points are not enough to make solid observations and form opinions. The data is currently too noisy.

So what to we do as we wait to have enough data points?

  1. We should unlock certain parts of the country where COVID cases are still low and the population density is lower than average. Areas such as rural communities and farms should not be under lockdown at the present moment.
  2. We should unlock certain sectors of the economy.
  3. Increase testing capacity
  4. Increase the capacity of the healthcare sector,

What do we do once we have enough data?

  1. We continue with social distancing measures to avoid huge spikes in cases. We have this thing under control, we must not loose control.
  2. Unlock low GDR areas and lock high GDR areas,
  3. Embark on massive testing, case isolation and home quarantines , especially in high GDR areas.

We play this game until a vaccine is available, at that point we will have the full economy operational and everybody back at work.

Identifying High and Low GDR Areas

A high GDR area is not necessarily the area with the highest number of cases. An area can have a high number of cases but with a rapidly declining GDR. Another area can have a low number of cases but with a rapidly increasing GDR.

Here is an example of a Ranking list that shows High Growth areas and Low Growth areas at a global level.

25 sec vid showing the DGRs ranking of countries with more than 10,000 cases . I have included South Africa for illustrative purpose, the country has only 3,000 cases. From this ranking , it can be observed that even though USA currently has the largest number of cases and largest daily increases in absolute numbers, COVID is getting under control there. Russia, Turkey and Brazil however have high GDRs , those are the next big hotspots.

The above can be stylized for provinces, cities and districts. The idea is to avoid a blanket approach at managing this thing. Only certain areas should be on full lockdown. The idea is also not to apply loose measures to the entire country. Only specific areas can have loose measures.

This “tighten-loosen” cycle applied using extensive data analytics such as Cluster Analysis will help country in balancing the need to have an open economy and the need save lives. It will allow the healthcare system to cope by avoiding extraordinary large spikes in all areas at the same time. This way, resources can be shared and transferred to the area that has the greatest need at a given point in time.

Epidemiologist say, eventually we will all get sick. However avoid a situation where all of us get sick at the same time and a situation where all of us are under lockdown at the same time. We should use DGRs to formulate the “tighten-loosen” policies as we go through the “dance phase”.

Ciao!

Financial Analyst, Cloud Accountant, Citizen Data Scientist, FPL Boss