By Vukosi Marivate, Elaine Nsoesie & Herkulaas MVE Combrink
COVID-19 is a unique event that has shaken the world. It has disrupted the way we live, how we work, and what we think. Across Africa, the arrival of COVID-19 also drew attention to the continent. We have had to live through grim forecasts of how “badly” the continent was going to respond to the virus, or whether the continent was different and we would not feel the impact. Given that we are still in the midst of the pandemic, we have a hard task of sifting through the opinions and reports to get to a better understanding of what has happened. We have to deal both with trying to better measure impact or contemplate if natural remedies would prevent spread. As data scientists, we believe that what is measured obscures shortcomings that otherwise might enlighten us on how we can better deal with such situations in the future.
Africa has significant experience dealing with infectious disease epidemics. For example, countries in West and Central Africa have responded over the decades to Ebola outbreaks, and Southern Africa has had HIV/AIDS to deal with. The experiences gained from these epidemics have prepared African health systems to respond to the pandemic. We are likely to see many research papers in the coming years dissecting what impact this preparedness may have had. In this article, we focus on how Africa worked to track COVID and what that might mean for data scientists in the future. What should we learn? Where did things go well? Where did things fail? How do we improve?
As the pandemic spread across the Northern Hemisphere, throughout the African continent questions formed about the potential impact of COVID-19 on different African countries. In many countries, COVID working groups were set up. These working groups were typically were made up of government and external experts who planned to look at different factors in the responses to COVID-19. In many instances, these groups used data to track COVID-19 and assist in modelling and data-driven decision making. One would have noticed the proliferation of country-led dashboards or infographics on the COVID-19 spread. In some countries, numbers were difficult to track and understand, because of low numbers of tests. The tracking of COVID-19 spread required a pipeline that could test, report, and aggregate information in a meaningful way for epidemiological and clinical surveillance.
We have seen international challenges to the free, transparent, and open reporting on the severity of COVID-19. Some African countries had these challenges as well, from denying the pandemic exists to refusing to release information on testing and confirmed cases. These challenges cannot be explained by simplistic reasons such as political pandering, but likely indicate challenges in resources available to respond to the pandemic. Countries have been stretched thin in a short period of time, and systems may not have the capacity to change direction this quickly. In this environment, how do you compile statistics and share meaningful information with both the public and policy stakeholders?
No one should underplay how COVID-19 will ultimately impact African countries. Its impact will not only be on healthcare; many sectors of society will likely be reeling from the sustained effects of the pandemic. There is already looming evidence about the adverse and secondary damage to other sectors such as education, crime, healthcare, and the economy. Decisions on border and business closures made during the early stages of the outbreak may also have lasting effects on countries in Africa.
COVID-19 has affected more than just health, and the effects will be with us for some time. As we move into second waves in some countries, we are now deciding how to rehabilitate economies, the education systems, and tourism. All of these decisions require data that crosses between national statistics offices and stakeholders. To better plan recoveries and interventions, organisations and states are working to use data to make choices about which interventions might be best. This process extends the need for data beyond the healthcare system toward a coordinated response driven by the public, private and non-governmental institutions. Data and data related issues are the ultimate reflection of people and capacity issues present within a system. If we are to combat negative outcomes, we should all work toward capacitating our nations to prepare for the future.
Counting is hard. It requires will, cooperation and resources that together improve policy. We need to learn how to set up the data infrastructure so that counting can catalyze data practices in the future. Yet, setting up a data infrastructure requires money and human capacity. Across the global population, we will have more emergencies to deal with. As such, governments must prepare adequately during the “peace times.” If we do not prepare, we will not get ahead to manage future crisis and crisis situations better. Investing in capacity and building the required skills to disseminate information in a more reliable way helps prepare us for the future. We should never sway away from training, innovation and incentivising education for the purpose of growth and improvement. Technical skills across all sectors—especially within healthcare—have served vital roles during the pandemic and will continues to do so. Capacitating the healthcare system with the technical skills to manage information, actively strive for excellence, and innovate still remains the foundation of preparedness, and drives the proactive strategies we need to be successful as a society.
Vukosi Marivate (https://dsfsi.github.io/) is the ABSA UP Chair of Data Science at the University of Pretoria. A large part of his work over the last few years has been in the intersection of Machine Learning and Natural Language Processing. Vukosi is interested in Data Science for Social Impact, and uses local challenges as a springboard for research. Vukosi is a co-founder of the Deep Learning Indaba, the largest Artificial Intelligence grassroots organisation on the African continent, aiming to strengthen African Machine Learning. He tweets at @vukosi.
Elaine Nsoesie is an Assistant Professor at the Boston University School of Public Health. She has a PhD in Computational Epidemiology, an MS in Statistics, and a BS in Mathematics. Her research is focused on the use of digital data and technology to improve health in global communities. Her work has also addressed bias in digital data. She is on the advisory boards of Data Science Africa and Data Science Nigeria. She is also the founder of Rethé (rethe.org), an initiative that provides scientific writing tools and resources to student communities in Africa to increase representation in scientific publications.
Herkulaas Michael Combrink is a medical biological scientist with more than six years data science experience with “Big Data” of institutional databases. Over the past seven years, he has been active in both healthcare and education. Herkulaas has won several awards for his work in Data Science, Data management and Healthcare. During the COVID-19 outbreak in the Free State, he has been seconded to assist the Free State Department of Health in data science and surveillance support. Additionally, Herkulaas is a PhD candidate in computer science at the University of Pretoria, South Africa.