Model scenarios
The NMG model (https://nmg-covid-19.sctechnology.co.za/) can be used to run off different scenarios for the COVID-19 pandemic in South Africa. The table below compares seven such scenarios. The scenarios all use the same starting point but have been calibrated using different basic reproduction numbers for the rate at which the disease spreads in a population. The calibrations have been done to show scenarios where the reported new infections peak earlier or later than the best-fit scenario.
Scenario 4 is the best-fit scenario for South Africa from the NMG model. This scenario is fitted to the confirmed cases to 5 April 2020 and assumes a disease progression in line with a basic reproduction number of 1.43. This means that 1.43 new cases are generated by each single infectious person. This scenario forecasts a peak in new infections of just over 1 million on 27 August 2020, and a total number of deaths from COVID-19 of 65000. At this rate of disease progression a forecast 53.5% of the South African population will get infected before herd immunity is reached and the virus stops spreading.
Scenarios 1, 2 and 3 show the effect of slowing down the rate at which the disease spreads. If South Africa is able to restrict the disease progression to a basic reproduction number of 1.21 by , say, imposing extremely restrictive lockdown conditions, then only 24.4% of the population gets infected. The problem comes when the the lockdown conditions are relaxed and the disease progression just starts up again.
Scenarios 5, 6 and 7 show the effect of speeding up the rate at which the disease spreads. A basic reproduction number of 2.97 (scenario 7) is equivalent to an environment with no lockdown or social distancing restrictions. At this rate of disease progression, 94.7% of the population will be infected with a peak of over 6 million new infections recorded on 27 May 2020, and some 115 000 deaths.
The NMG model is one of a class of SEIR models. Other models have been fitted to the COVID-19 pandemic that give very different forecasts. Boston Consulting Group uses a model that forecasts the pandemic for South Africa reaching a peak in new infections on 1 June 2020 at a level of circa 50000 infections reported on that day. I will look at this in more detail in the blog tomorrow. For now, I mention this just to stress how different models can give widely different results.
7 April 2020
Scenarios 1, 2 and 3 show the effect of slowing down the rate at which the disease spreads. If South Africa is able to restrict the disease progression to a basic reproduction number of 1.21 by , say, imposing extremely restrictive lockdown conditions, then only 24.4% of the population gets infected. The problem comes when the the lockdown conditions are relaxed and the disease progression just starts up again.
Scenarios 5, 6 and 7 show the effect of speeding up the rate at which the disease spreads. A basic reproduction number of 2.97 (scenario 7) is equivalent to an environment with no lockdown or social distancing restrictions. At this rate of disease progression, 94.7% of the population will be infected with a peak of over 6 million new infections recorded on 27 May 2020, and some 115 000 deaths.
The NMG model is one of a class of SEIR models. Other models have been fitted to the COVID-19 pandemic that give very different forecasts. Boston Consulting Group uses a model that forecasts the pandemic for South Africa reaching a peak in new infections on 1 June 2020 at a level of circa 50000 infections reported on that day. I will look at this in more detail in the blog tomorrow. For now, I mention this just to stress how different models can give widely different results.
7 April 2020
Comments
Post a Comment