Visualising the cross-country trends of a global crisis and lessons for Kenya on reopening and reclosure

The Big Question

To what extent have the COVID-19 containment strategies adopted in different countries been successful? Which country can claim long-term success worth emulating yet? No! It must be still too early for a novel virus sending scientists to their studious research corners.

Then, what lessons can Africa draw from the global COVID-19 developments? As a key reference point in East Africa, should Kenya be open to alternative views or just read the signs from other countries and resort to tougher lockdown measures since the easing of restrictions on July 6, 2020, subsequently saw a substantial surge in confirmed COVID-19 cases?

This special edition consolidates the key lessons from the previously shared IBD models of COVID-19 and generates a compact bandwidth of projected scenarios of growth in COVID-19 cases at various scales: globally, in Africa, and nationally in Kenya and several top countries in the COVID-19 list such as the USA, Brazil, India, Russia, Peru, and Chile. The objective is to contribute to the generative dialogue needed to inform long-term policy, planning, and monitoring of the pandemic and its impacts on economies.

The Reopening Dilemma

Read et al. (2020) used a transmission model at the time the novel coronavirus began spreading fast in China to estimate a basic reproductive number, R0, of 3.11. This factor ranks COVID-19 as a highly infectious disease. They further estimated that 58–76% of transmissions must be prevented to stop the surge.

The World Health Organization (WHO) had by July 2020 published several situation reports confirming that community transmission had been the established mode of spread of the novel coronavirus, even across the latecomer in COVID-19 cases which was the African continent. Testing capacity, case fatality rates, and recovery rates constitute the triangle of parameters used to rate the relative performance of countries in the fight against the pandemic. Technologically leveraged location-based contact tracing generates the actionable spatial intelligence required to contain the spread COVID-19.

Probing questions querying if country reopening schedules have been hurried are gaining currency and prominence in popular discourse. Countries such as Israel and South Korea, after recording remarkable success in containing the spread of the pandemic, have experienced a significant resurgence of cases.

In Kenya, there has been a surge in cases since the easing of movement restrictions on July 6, 2020. Fear is growing that places of essential services such as hospitals, government offices, courts, among others could turn out to be a nexus of hotspots for rapid transmission of the novel coronavirus. A compact bandwidth of model scenarios is needed to inform policy and planning in such situations, including the possibility of helping to modify popular behaviour to contain the pandemic at scale.

A Snapshot of Telling Trends: Global, Top Countries, Africa, and Kenya

Models should be as simple as possible yet as complex as necessary, noting that all models are abstractions of reality — hence generally wrong and can only be fit-for-purpose tools (Adero and Kiema, 2011; Bellinger and Fortmann-Roe, 2013). System dynamics models and agent-based models have been widely applied. A data-driven mathematical modelling approach was, however, this study’s economical choice as informed by the abundance of COVID-19 data already generated globally to the country level, enough to support compact simulations.

By July 22, the global recovery rate was 61%, higher than Africa’s 55% by the same date. The case fatality rate in Africa of 2% was half the global average of 4%. This independent mathematical modelling of the trend of the growing data on COVID-19 cases yielded the following projections.

• Globally, the total cases had hit 13 million by July 13 with a case fatality rate of 4% and a recovery rate of 58%. The observed 1.9% daily exponential rate gave a simulation model that predicted about 14 million total cases by July 15, 15 million cases by July 19, 17 million by July 25, and about 19 million cases by August 01, 2020. The peak of this upward global COVID-19 curve was by then not in sight yet.
• The USA has been leading globally in total cases, crossing the 4 million mark on July 22 with a case fatality rate of 4% and a recovery rate of 47%. Based on the same data-driven mathematical models IBD has been applying, the projections for the USA came to 4 million cases by July 22 (which was realised) and 4.7 million cases by July 31, 2020. The curve of the USA was still rising with the peak not in sight by then.
• Brazil has recently been second globally in total COVID-19 cases. Starting May 25, 2020, Brazil has been on a decisively upward curve with the peak not in sight by July 23. The case fatality rate and recovery rate by July 13were 3.9% and 65%, respectively. The IBD model made for Brazil, a polynomial of gradient 431.6x + 20885, has predicted 2.5 million cases by July 26, 2.7 million by July 31, and to hit 3 million cases by August 5, 2020.
• Russia has been ranking fourth globally, after India. By July 13, the country had a low case fatality rate of 1.6% and a recovery rate of 69%, with over 730,000 total cases. Russia’s curve has shown signs from June 18, 2020, of a reduced daily growth rate of 1% in reported COVID-19 cases. The model projects for Russia between 821,162 and 864,415 cases by July 28 and between 842,818 and 901,493 cases by August 1, 2020 (Figure 1). Russia’s simulation curve has since June 18 been gently facing downward with a gradient of 7430–48x, which if unchanged, theoretically peaks at 1,137,412 cases on November 20, 2020. Over this study period, Russia was the first case of the top-five countries battered by COVID-19 showing a foreseeable peak in COVID-19 cases (see Figure 1).

• By July 13, India had almost 900,000 cases with a case fatality rate of 2.6% and a recovery rate of 63%. It had started showing a growth rate in confirmed COVID-19 cases that had slowed a bit from the 4.15% daily exponential rate of late May to mid-June, to a lower rate of 3.3%. India’s curve has also been on an upward rise with the peak not in sight by July 23, 2020. The projection bandwidth for India predicted a million total cases by July 17 and between 1.4 and 1.6 million total cases by July 31.
• Peru, ranking sixth globally by July 22, has recorded a high number of cases for its population of 33 million, about 330,000 cases by July 13. The trend studied since June 19 projected its total cases to about 390,000 by July 31. By July 22, 2020, Peru’s case fatality rate was about 4% with a recovery rate of 70%.
• Chile was projected by the model to record about 340,000 COVID-19 cases by July 31, a projection that is likely to be exceeded by actual cases. In late July, the country featured a high recovery rate of about 90% with a case fatality rate of about 3%.
• Africa is likely to record between 815,444 and 1,013,069 total cases by July 28 and between 871,542 and 1,118,495 total cases by July 31, 2020. The COVID-19 curve for Africa was still rising decisively with the peak not in sight by July 23, the lower trajectory being a polynomial of gradient 232.8x + 8109 and the upper range of projections assuming a daily exponential growth rate estimated at 3.3% since June 14, down from the former rate of 4.16% (Figure 2). By July 22, Uganda, Seychelles, and Eritrea were still the only African countries which had not recorded any COVID-19 deaths.

• South Africa, ranking fifth globally in July, is likely to record between 600,000 and 700,000 cases by August 1, 2020, going by the model. The COVID-19 curve for South Africa has been rising at an exponential rate with the peak not in sight by July 23, 2020. South Africa’s daily growth rates have been between 4% and 5% over the period April 15 — July 16, 2020.
• In East Africa, Kenya started realising a surge in confirmed COVID-19 cases in May 2020 as the country increased her testing capacity. The model developed here for Kenya has used past data to arrive at a compact bandwidth of three simulated scenarios. The simulated scenarios have since May 24 been keeping the actual COVID-19 curve in between the lower and upper projected trajectory, as shown in Figure 3. The green line is the modelled trend of the growth in cases before the easing of movement restrictions by Kenya’s President on July 6, 2020, to allow for free movements and economic activities across the vast nation of some 50 million people. It can be seen that the cases surged within two days after July 6, 2020, the black line of actual cases veering off the green trajectory to match the yellow line simulating the new growth rates. As the latest trend of the black line of confirmed cases shows in Figure 3, Kenya is likely to record more than 18,500 cases by July 28, 21 days after the easing of nationwide movements. Ceteris paribus, the model projection shows a high likelihood of Kenya exceeding 20,000 confirmed cases by July 31. Assuming no major limiting constraints on Kenya’s testing capacity, the model projects that the confirmed cases could number between 70,000 and 110,000 by September, then between 700,000 and 1.7 million by November 2020.

Comparative Metrics on Fatality, Recovery and Testing Rates

There is a growing body of evidence that these three rates hold an important key to understanding the trends of the pandemic. For a simplified visualisation and comparison of cross-country performance, the analyses have been presented in radar charts. Because of the wide variation in the normalised testing indices for different countries, as explained below, a logarithmic scale (base 10) has been applied to graphically scale the normalised testing rates. Mauritius, for example, has established a daily normalised testing index of more than 1,000, far ahead of the double-digit scores for most African countries.

In Europe, Russia (less than 2%) and Germany (4%) have been the focal points of low case fatality rates. Higher case fatality rates in Europe have been observed in France (17%), the UK (15%), and Italy (14%) — current as at July 22, 2020 (Figure 4). As shown by the gap between the black and red lines, the European countries in this group have realised an increase in case fatality rate, significantly over the lower the rates they recorded by March 31, 2020.

Recovery rates have been taking time to improve, exemplified in Africa by the case of Ghana whose recovery rates remained below 33% up to late May before jumping to over 70% by the end of June and over 84% by mid-July, 2020. In modelling, such lagged effects are common, captured as delays in system dynamics models. Global recovery rates have been peaking at 90–97% for the countries that have realised high steady rates over time, such as Mauritius, China, and Germany (Figure 5).

Key Lesson 1:

The cross-country COVID-19 analyses at different times have confirmed that African countries have generally realised lower testing rates, lower case fatality rates, and lower recovery rates than in Europe and other parts of the world. The delay in realising recoveries imply that African countries, as latecomers in the COVID-19 statistics, may well still have high chances of realising more recoveries to increase their scores. The lower case fatality rates across Africa are so far encouraging, seemingly lending early credence to the hypothesis advanced in favour of Africa’s youthful demographic, with a median age of only 20 as opposed to over 40 in Europe (see Njenga et al., 2020).

The observation across the board is that the countries that have been testing their populations to a greater depth in Africa have been posting progressive performance metrics (Figure 6). To reflect the vast differences in population sizes (e.g. Nigeria’s over 200 million against only 1.2 million for Mauritius), the daily COVID-19 testing rates have been normalised into measures per million people in the respective countries. In Kenya, the daily population-normalised testing rates have been increasing slowly but steadily from a low of 3 tests per million people per day in March to 36 tests per million people per day by July 23, 2020.

Key Lesson 2:

Enhancing the testing capacity in Kenya and other African countries is justified, being a key source of the actionable intelligence needed to inform effective case management and the required threshold of national and county preparedness.

Key Lesson 3:

Country population sizes matter when comparing COVID-19 testing rates and their implications for containment policy and strategies. India has an interesting combination of high confirmed COVID-19 numbers and a relatively low testing index for the top country cases, implying that India’s cases can explode over time with increasing testing rates. The county’s leading population size of 1.3 billion is playing a key role in the equation of the projected surge in COVID-19 cases.

Conclusion and Recommendation

Generally, the peaks of the COVID-19 curves for the countries studied here were not in sight by July 23. The implication of this scenario for country leaders is to continue exercising caution because COVID-19 cases are still on the rise globally. Any phased reopening must be informed by well-researched evidence and a keen consideration of the behaviour of the citizenry.

Kenya eased restrictions on movement across the country on July 6, 2020, a move that has earned criticism and praise in equal measure. The model of scenarios shows that the confirmed COVID-19 cases started surging within 48 hours after the easing of restrictions. The model projection has shown a high likelihood of exceeding 20,000 cases by July 31. Ceteris paribus, the model projections have shown a high likelihood of recording confirmed cases numbering between 70,000 and 110,000 by September, then between 700,000 and 1.7 million cases by November 2020.

In the face of the COVID-19 pandemic, nations are trapped in a complex and confounding web of health, economic, and political choices. The many unsettled questions on the nature of the novel coronavirus mean that every country is on a learning curve dotted with uncertainties, risks, and pitfalls. No country can claim any absolute and lasting triumph over the disease yet. Adaptive management of the pandemic must take centre stage.

After recording exemplary performance in containing the spread of COVID-19 and reopening, countries such as South Korea, Singapore, the UK, the USA, among others have had to effect reclosures in various regions as cases surge again with the easing of restrictions. Israel, exemplary among the developed economies in containing the pandemic with a case fatality rate of less than 1% and a recovery rate of 42% as at July 22, has lately been experiencing another spike in cases. This is a reminder of the sharp exponential rise of 21% this modelling series estimated for Israel in the second half of March 2020.

Remaining open to new ideas from all quarters and re-examining any hardliner positions that stand in the way of progressive change is the wise position leaders should take. Every citizen, however, holds the individual responsibility of discipline and precaution to stem a COVID-19 crisis.

Country leaders can do justice to COVID-19 research outputs by making balanced and well-advised decisions. Their COVID-19 speeches must not forget to place emphasis on behaviour change agency and urgency, discipline, and a judicious balance between competing spontaneous ideas and concrete scientific ideals.

Finally, every country is facing limitations in containing COVID-19. However, if you keep defending your limitations, you will have to keep and maintain them. Country leaders must act and choose allowable COVID-19 tolerances for their countries. Wise application of tolerances is common in applied precision sciences such as surveying. Though tolerance levels are all fit-for-purpose vehicles, not all of them fit all purposes. To match the uniqueness of the challenges at hand, wise choice must prevail. That is the key challenge of crisis leadership at the moment.

References

Adero, N.J. and Kiema, J.B.K. (2011). Flow-based Structural Modelling and Dynamic Simulation of Lake Water Levels. In: T. Gasmelseid (ed), Handbook of Research on Hydroinformatics: Technologies, Theories and Applications, 316–331; and Information Resources Management Association (IRMA) (ed), Green Technologies: Concepts, Methodologies, Tools and Applications, 798–814. USA: IGI Global.

African Arguments (2020). Coronavirus in Africa Tracker: How many covid-19 cases & where? Africanarguments.com.

Bellinger, G. and Fortmann-Roe, S. (2013). Beyond Connecting the Dots: Modelling for Meaningful Results. eBook available online.

Impact Borderless Digital (IBD) 2020. IBD research and publication series. Available from: impactborderlessdigital.com

Ministry of Health, Kenya (2020). Daily COVID-19 data briefs in Kenya.

Njenga, M.K., Dawa, J., Nanyingi, M., Gachohi, J., Ngere, I., Letko, M., Otieno, C. F., Gunn, B.M. and Osoro, E., Am. J. Trop. Med. Hyg., 00(0), 2020, 1–6, doi:10.4269/ajtmh.20–0474.

Read, J.M., Bridgen, J. RE., Cummings, D. AT., Ho, A., Jewell, C.P. (2020). Novel coronavirus 2019-nCoV: early estimation of epidemiological parameters and epidemic predictions. COVID-19 SARS-CoV-2 preprints from medRxiv and bioRxiv. doi: https://doi.org/10.1101/2020.01.23.20018549

World Health Organization (WHO) 2020. Novel Coronavirus (2019-nCoV) situation reports, January — July 2020. Accessible from: https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports/

Worldometer (2020). COVID-19 Coronavirus Pandemic. Statistics accessible from https://www.worldometers.info/coronavirus/