MS&E 433 Mini-Project: Modeling for Botswana, April 8-April 28

Team
Brief problem statement
Useful resources
Findings

TEAM


BRIEF PROBLEM STATEMENT

Currently, there is little modeling and forecasting on how Covid-19 is impacting African nations. This is due, in part, to an incongruence between the presently available data and that needed by models. Starting with Botswana and Rwanda, we will gather necessary data and work to adapt it for publicly available models. In particular, we will use publicly available data and leverage our connections in both countries to source any other missing data for the Covid-19 Scenarios Model. The team behind the Covid-19 Scenarios Model has identified both hospital resource data (e.g. number of ICU beds) and age distributions as critical to the proper functioning of the model, so we will pursue these two directions first. If we are able to validate our process, we hope to expand the effort to include other, strategically chosen countries within the continent.

USEFUL RESOURCES

Covid-19 Scenarios Model

Johns Hopkins COVID-19 Dashboard

Worldometer Coronavirus Dashboard

COVID-19 Botswana Dashboard by the Botswana International University of Science and Technology

COVID-19 Botswana Dashboard by the University of Botswana for the Ministry of Health and Wellness

World Health Organization COVID-19 Situation Reports

PROGRESS UPDATE

As stated in the above Problem Statement, we set out to find hospital resource data and age distribution data for the Covid-19 Scenarios. Other collaborators on the Covid-19 Scenarios project were able to locate the age distribution data on the CIA WorldFactbook and linked it to the Covid-19 Scenarios. We scaled their work by writing a Python program that parses out the age distribution data for any country from the CIA World Factbook and stores it in as a usable format for models requiring such distributional data. The parser may be found on our GitHub. We hope this will help bridge the gap between modeling efforts and the challenge of scarcity of relevant data on developing countries.

Unlike population data, hospital resource data has proven hard to find online. Efforts to establish connections with the Ministry of Health in Rwanda has been slow-going. We have had better luck in Botswana, where we have been able to establish contact with researchers at the Botswana International University of Science and Technology, the clinical response team at the Sir Ketumile Masire Teaching Hospital of the University of Botswana, and officials at the Ministry of Health and Wellness. Even with these connections, we still do not have the hospital resource data due to long communication cycles. We remain optimistic that we will receive this data and be able to leverage it for our continued modeling efforts for Botswana.

Our conversations with the response team and updates published by the COVID-19 Presidential Task Force in Botswana indicate that the authorities in Botswana have managed to contain the spread of COVID-19. The priority is shifting towards easing restrictions and reopening the economy, while ensuring there is no resurgence of cases. With the arrival of winter and the flu season, the clinical response team will need a testing protocol to allocate the limited coronavirus tests. Additional issues arise with the knowledge that, currently, the country does not issue tests for the seasonal flu, and this might confound efforts in identifying COVID-19 cases. For the second phase of our project, we plan to continue our current efforts and expand into aiding the Botswana Task Force’s goals.