CoSimPol: New Covid-19 Simulator for Policy Makers

A new, system dynamics based simulator, Covid-19 Simulator for Policy Makers - CoSimPol has been released. It can be calibrated to any country or local context. 

The on-line simulator may be found here: https://exchange.iseesystems.com/public/shamim/cosimpol/index.html#page1

It includes the following fully user configurable operating parameters:

• Number of known cases on the first day of the simulation.
• Assumptions on proportions of unknown cases to known cases.
• Hospital and critical care bed capacity, as well as building new bed capacity e.g. field hospitals.
• Quality of in-hospital care.
• The possibility to model up to two lockdowns of varying intensity.
• Non Pharmaceutical Interventions (NPI) with a non-linear decay of effectiveness.
• Modelling the possibility of “losing” immunity.
• Three different vulnerabilities of infected patients (non-hospitalisation, hospitalisation and critical care).
• Modelling different values of R0’.
• All of the key assumptions are available to the user to critique, and most may be changed by the user in order to validate the results.
• A full sheet of assumptions for the base calibration is shown.

The base case has been for calibrated for South Africa, with an initial R0’ of 2,94. The salient points from the base case is as follows:

1.  Infections will peak towards the end of June 2020
2. The number of deaths will exceed 174 000
3. Lockdowns in and of themselves do not “flatten” the curve significantly although they do reduce the peak infections.
4. The major effect of lockdowns is to delay the peak, time-shifting the spread of the epidemic into a later time, and reduces the number of infected people relative to no lockdown.
5. This may enable the country time to build new bed and critical care capacity.
6. Lockdowns if applied inappropriately may be harmful.
7. The timing and duration of a lockdown is critical on reducing the number of fatalities.
8. Given the level of R0' in the base case there is a level of natural "herd immunity" without a vaccine being available.
9. In addition to reducing peak infections by appropriate policy interventions, it is also possible to reduce the absolute number of infections, and hence a form of herd immunity.

The above is based on a few scenarios. There are a range of possibilities given the complexity of the underlying systemic factors related to the pandemic. This will require much more intense sensitivity analyses to arrive at a better understanding of the pandemic, intervention strategies and impacts.

The model does not factor in economic and social impacts yet, but could in principle be extended for such a purpose. In a similar vein, it could be extended to explore a range of outcomes and impacts if a vaccine is discovered.

Three training webinars on CoSimPol are available here:

• https://www.youtube.com/watch?v=e6jBHMScYqk
• https://www.youtube.com/watch?v=co2_A-XPL40
• https://www.youtube.com/watch?v=-_Ugjxf11Sw

Article updated 10 May 2020