Georgia Recovery Model
Georgia initially had a super-spreader event occurred on Feb. 29, at the funeral of the beloved Andrew Jerome Mitchell. The daily mortality rate accelerated in the ensuing 3rd and 4th weeks after the funeral, which is what we would expect. This drove the initial ramp in deaths. As the infection spread to the Atlanta area it grew slowly and was largely suppressed by the response in that area, starting with a state of emergency declared on March 15. The death rate is less than that of a typical flu season (Pneumonia and Flu). A small percentage of the overall population has been exposed to the virus and the reproductive rate has been pushed well below 1.
In or simulation we apply 2 months of each phase of the White House / CDC plan for recovery. Our analysis a derivation of simulation input parameters is here. We allow for mild summer degradation in the transmission rate and mortality of the virus. Note that once a return to normal occurs the infections spike up and the peak is greater than the initial infection which was tamped down early.
This scenario is is just to illustrate that recovery measures can be implemented now with minimal risk, and buy much needed time for the development of therapy measures. The decision to transition to full recovery can be deferred and decisions can be based on data (active test data to track infection trends, tracking hospitalization rates and adjusting mitigation strategy to attack high infection rate opportunities (like large events) and protect those who are vulnerable.
We have assumed that 50% of cases are asymptomatic. If the actual asymptomatic rate is higher, then the return infection rate will be diminished.
Dynamic input parameters are below showing the phased response.
A population of 2.5 Million was used in the base simulation:
Below is the cumulative numbers from this simulation. The number of infected people in the general population is fairly low, and as such, Georgia is vulnerable to a second wave as our analysis shows, but will reach herd immunity once 40% of the population is exposed. Most of the deaths are within the general population. The death rate is pretty low in this scenario. 6500 deaths divided by a population of 4 million infected (dead+recovered+infected) is a death rate close to 0.16%. If a higher percentage of infected people are asymptomatic, this rate would be lower and approach 0.1%.
The strategy of protecting the vulnerable results in 85% of them being protected while herd immunity is established.
The reproduction factor over time is well below 1 for Phase 1 and Phase 2 but climbs in Phase 3 as the number contacts increases and the climate suppression of the summer diminishes. Note that the death rate from this slow growth is delayed by 3 months to onset. With effective testing and data monitoring this growth in multiplication factor can be detected and adjustments can be made if necessary to public policy to reduce high infection risk activities.
The demand on hospital resources was very light in Georgia in general. The numbers below are per 10 Million population. The second wave we see in our simulation is significantly higher for standard hospitalizations and should be considered. Georgia has around 22,000 hospital beds so this second wave would be about 50% of that capacity. The critical care demand is much lower, about twice the first wave, primarily because we have protected the vulnerable from the outbreak. This load is likely to be manageable. If effective therapies are developed to reduce death rates and improve recovery times, this load will be significantly reduced.
Additional input parameters are shown below: