Data scientists angry over the “perversion of science” used to justify South Africa’s lockdown

Pandemics Data and Analytics (PANDA), a group of data scientists which has repeatedly criticised the South African government’s approach to the national COVID-19 lockdown, has explained why it is so frustrated with the government’s modelling.

In a statement published to its website, PANDA said that based on its assessment of the COVID-19 pandemic’s spread in the country and the economic impact of the lockdown, the total loss of life caused by lockdown would dwarf even the most optimistic assumption of lives saved from COVID-19.

“South Africa is on the way to a sovereign debt crisis and all that entails,” PANDA said. “In our context, the economy was essentially too weak to weather a lockdown to start with.”

Despite this, South Africa is still in lockdown, and PANDA said its frustration stems from the flawed science on which the government is basing its decisions.

This science comprises models which predicted up to 351,000 COVID-19-related deaths in South Africa, which PANDA said was overblown and used as the rationale for one of the harshest lockdowns in the world.

“From the start, we highlighted concerns with those models, and how they failed to explain real-world observations from countries ahead of us on the curve, but that fell on deaf ears,” PANDA said.

“We mentioned that COVID-19 deaths in South Africa would most likely land in the region of 10,000 to 20,000. All the current data confirms this estimate.”

PANDA said that despite its models proving far more accurate than the projections on which the government based its lockdown response, the organisation’s input has been regularly denied.

The organisation said it is frustrated that the government and the modellers who supply the NCCC with projections will not consider these more accurate models, as the current lockdown is based on bad modelling and is unnecessarily harmful.

“We take no pleasure in being proven right in retrospect while millions of our fellow South Africans go to bed hungry, and the futures of millions of children are put in jeopardy by the repeated closures of schools,” PANDA said.

Active cases plateauing

Far from reaching the feared COVID-19 related death figures of more than 300,000 stated in early predictions, the number of active COVID-19 cases in South Africa seems to be plateauing.

The number of active cases over the past week has been holding steadily between 160,000 and 170,000.

These active case figures reflect a nationwide epidemic peak, although the number of active cases is many municipalities is continuing to increase rapidly.

The amount of testing has also declined slightly over the past week, which has resulted in the number of recorded tests each day being lower than normal.

Data published by the National Institute of Communicable Diseases of South Africa (NICD) is collated by MediaHack and represented via an online coronavirus dashboard which outlines the trends in active and daily new cases across provinces and the country as a whole.

This data conforms closely with PANDA’s projections, reflecting a reproduction number close to 1.0 nationwide and lower than 1.0 across a number of provinces, meaning the virus may be beginning to be contained in certain areas.

“Perversion of science”

PANDA has said that it has attempted to engage with the modellers who provide projections to the government.

“In May we were invited to assist, in a limited capacity, in advising the Western Cape government. We quickly pointed out, with confidence, that the model projections of the hospital bed and ICU demand for the province were grossly overstated,” PANDA said.

“However, this put the decision-makers in a very difficult position. While an official model exists that says otherwise, they simply cannot take the risk of trusting PANDA and being caught with their pants down.”

It is clear now that the Western Cape has passed its peak, but rather than use this as an opportunity to revise their projections, PANDA said the modellers have doubled down on their projections.

“We admit that we cracked and made some of the modellers and academics the ‘face’ of this tragedy,” PANDA said.

“We did this because we perceive them as complicit, but perhaps unfairly reached the conclusion that this perversion of science was deliberate and malicious.”

“In the face of mounting real-world data, the modellers keep feeding the fear and feeding our government the science it needs to justify their reaction as rational and maintain the crippling lockdown,” PANDA said.

The graphs below show the cumulative and active cases across the country, the daily new cases in each province, and the current reproduction estimate for the country.


Cases vs Active Cases in South Africa

Cumulative cases vs active cases 29 July 2020


Daily New Cases in Each Province

Daily new cases in each province 29 July


Reproduction estimates for South Africa

Rt estimates South Africa 26 july

Now read: Curve of active COVID-19 cases in South Africa is flattening

Latest news

Partner Content

Show comments

Recommended

Share this article
Data scientists angry over the “perversion of science” used to justify South Africa’s lockdown