Hello:

Here are a “short” summary and action items from the meeting:

Present: Prof. Simon Connell, Dr. Somiealo Azote, Sister Mary Taabu, Cyrille Haliya, Toivo Mabote, Kondwani Mwale, George Zimba, Ssebandeke John, Kétévi.

Benize Niyikiza could not connect due to limited internet access but she sent updated results.

We discussed the SIR Model fits to the various data and concluded that we need to improve the model in order to improve the fits:

1. First, we will move from the SIR model to the SEIR model by introducing the dE/dt differential equation to describe the “Exposed by not infectious population”. For the next meeting, everyone should introduce this modification and re-make the plots. We want to make sure that we all get the same results before extending the model further. The equation for dE/dt can be found in this paper https://arxiv.org/pdf/2004.07208.pdf Equation (1).  When you introduce dE/dt, please make sure to update dI/dt as well. Basically, the new model is SEIR, described by Equation (1) in that paper;
2. Sister Mary will find out and let us know about the conditions, requirements and scope for the grant application mentioned here https://www.icgeb.org/activities/grants/. The deadline is May 15. If we are eligible to apply, we will prepare 1 joint application;
3. We will meet again on Thursday April 30 at 14:30 GMT. At this meeting, we will discuss 2 things:
   1. Everyone’s implementation of the SEIR Model according to Equation (1) in the paper mentioned above;
   2. Prof. Connell will tell about how we might evolve the parameter b with time; Also how we could modify the equations with added terms to fit the data better.
4. After our meeting on April 30, we will move the meetings to Fridays at 14:00 GMT from May 8, to be joint with the ones organized by Prof. Connell.
5. When you move from SIR to the SEIR, please note:
   1. b is the same b that we were using before in the SIR;
   2. g is the same parameter k that we were using before the SIR;
   3. The SEIR has another parameter a related to the incubation period. Assuming the average incubation period is 7 days, then a=1/7.
   4. Work with the fractional distributions s= S/N, e=E/N, i=I/N and r = R/N with the requirements that ds/dt+de/dt+di/dt+dr/dt=0; s(0) =1; e(0) = 0; i(0)=1.25e-6; r(0)=0.
6. We decided that we stay with the differential equation model and try to improve them instead of adopting the integration equation model described in the paper mention in 1) above.
7. For the people that asked for Kétévi c++ code, he put it on the agenda page. This code requirement the ROOT library.

Regards, Kétévi.