Thread by @pieterstreicher

1 of 4:

8 December Gauteng #Omicron short term projections:

Cases, hospitalisations and ICU beds are all running below my projections.

Ventilated beds are slightly above.

2 of 4:

The method is simple.

1. Determine the daily growth rate during the exponential phase for all variables.
(cases 30%, hosp 12%)

2. Estimate when cases will peak based on an Rt extrapolation. Alternatively, fit a Gompertz curve to the case trajectory.
(12 Dec)

3 of 4.

3. Apply a consistent drop in growth rate from the point when growth is not exponential anymore to the point when Rt=1.0. (Alt. - Gompertz fit)

4. Add a suitable lag period for the hospital variables. I used 6 days for hospital and another 5 days for ICU/vent beds.

4 of 4:

These are the actual numbers.

Growth rates could fluctuate due to reporting delays.

So far the projections are reasonable, with perhaps a possibility of cases peaking earlier.

update:

8 December 7day average cases added for Gauteng:

9,144 (9% growth from previous day)

This is significantly lower compared to the peak of the exponential phase (28% growth per day).