key: cord-0990369-g6pikj1r
authors: Christophe Dongmo, F. M.; Monique, A. N.; Amandus, A.; Momah, K. V.; Steve, M. F.; Simeon Pierre, C.
title: Does sub-Saharan Africa truly defy the forecasts of the COVID-19 pandemic? Response from population data
date: 2020-07-07
journal: nan
DOI: 10.1101/2020.07.06.20147124
sha: a2fb77d80bb523f8708f0e7e36efb6dd33cc218b
doc_id: 990369
cord_uid: g6pikj1r

Introduction. Since its identification, the COVID-19 infection has caused substantial mortality and morbidity worldwide, but sub-Saharan Africa seems to defy the predictions. We aimed to verify this hypothesis using strong statistical methods. Methods. We conducted a cross-sectional study comparing the projected and actual numbers as well as population proportions of COVID-19 cases in the 46 sub-Saharan African countries on May 1st, May 29th (4 weeks later) and June 26th (8 weeks later). The source of the projected number of cases was a publication by scientists from the Center for Mathematical Modeling of Infectious Diseases of the London School of Hygiene & Tropical Medicine, whereas the actual number of cases was obtained from the WHO situation reports. We calculated the percentage difference between the projected and actual numbers of cases per country. Further, N-1 chi-square tests with Bonferroni correction were used to compare the projected and actual population proportion of COVID-19 cases, along with the 95% confidence interval of the difference between these population proportions. All statistical tests were 2-sided, with 0.05 used as threshold for statistical significance. Results. On May 1st, May 29th and June 26th, respectively 40 (86.95%), 45 (97.82%) and 41 (89.13%) of the sub-Saharan African countries reported a number of confirmed cases that was lower than the predicted number of 1000 cases for May 1st and 10000 for both May 29th and June 26th. At these dates, the population proportions of confirmed Covid-19 cases were significantly lower (p-value <0.05) than the projected proportions of cases. Across all these dates, South-Africa always exceeded the predicted number and population proportion of COVID-19 infections. Conclusion. Sub-Saharan African countries did defy the dire predictions of the COVID-19 burden. Preventive measures should be further enforced to preserve this positive outcome.

On December 12 th , 2019, an epidemic caused by the coronavirus disease 2019 (COVID- 19) emerged in Wuhan, China [1] . The epidemic developed rapidly and spread to many countries. On March 11 th , 2020, the World Health Organization (WHO) declared the COVID-19 outbreak a pandemic [2] .

The first case of COVID-19 on the African continent was detected in Egypt on February 12 th , 2020 [3, 4] . Since then, the pandemic has substantially progressed and as of July 4 th , 2020, African countries reported a total of 342 415 confirmed cases and 6 628 deaths [5] . The pandemic did catch the continent off guard and has since obliged political leaders to take important decisions constantly. The rapid progression of the sanitary crisis constrains those stakeholders to make these decisions based on information of the highest accuracy possible, with regards to the number of lives at stake. Since history is the best teacher, historic data are usually utilized as reference for decision making purposes. However, COVID-19 is a novel disease and thus no historic data exist on the pandemic. Therefore, to raise alarms and build public health strategies, public health officials can only rely on predictive models. These models are informed by the existing knowledge about population features and habits as well as the growing evidence on the COVID-19 infection.

Sub-Saharan African populations are very dense and one of the key characteristics of their social life is the presence of quasi-constant human contacts (handshakes, hugs, accolades), which make it very difficult to apply social distancing. Furthermore, potable water is not readily available, and poverty prevents most of these populations from affording either the face masks or the materials necessary to implement and enforce ideal hygiene sanitation measures. Finally, a great percentage of sub-Saharan African subjects rely on a daily income to sustain a living; as such, it is not possible to maintain confinement measures for long periods of time. Based on these facts, dire predictions were made about the evolution of the COVID-19 infection in Africa. A statistical model built by Pearson et al. (2020) , from the Center for Mathematical Modeling of Infectious Diseases of the London School of Hygiene & Tropical Medicine (LSHTM) predicted that most African countries would report 1000 cases by May 1 st and 10,000 cases few weeks later [6] .

The World Health Organization alerted on April 17 th , 2020 that Africa would become the next epicenter of the COVID-19 pandemic [7, 8] . As the pandemic has evolved, several press releases have highlighted that sub-Saharan African countries seemed to have defied the odds of such predictive models [9] [10] [11] . Although seemingly positive, such news have left most researchers dissatisfied because there has not been any scientific appraisal of the difference between these predictions and the actual data. In effect, to the best of our knowledge, there exists no quantitative assessment of the difference between the actual and projected number of COVID-19 cases in sub-Saharan Africa. Further, no studies have attempted to determine if there is a statistically significant difference between the projected and actual prevalence of the COVID-19 infection and explain any pattern. Our research therefore aimed to use the predictions made by Pearson et al. (2020) , to (1) determine the percent difference between the actual and projected number of COVID-19 cases in sub-Saharan African countries at the indicated dates, (2) assess if there is a statistically significant difference between the projected and actual population proportion of COVID-19 at these dates and (3) discuss any pattern.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted July 7, 2020. . https://doi.org/10.1101/2020.07.06.20147124 doi: medRxiv preprint

We conducted a cross-sectional study analyzing the projected and actual number as well as population proportion of COVID-19 cases in the 46 countries of sub-Saharan Africa. The date of May 1 st was the date indicated by Pearson et al. (2020) , [6] to describe the day at which, on average, most African countries were projected to have reached 1000 cases. They further stated that most African countries would have reached 10000 cases "couple of weeks" after May 1 st . Because the timeline to reach the 10000 cases was not specifically indicated, we choose to make comparisons with the actual number of confirmed cases reported on May 29 th (4 weeks after May 1 st ) and June 26 th (8 weeks after May 1 st ). Pearson et al. (2020) , obtained these predictions from robust statistical models built around established epidemiologic parameters like the average number of additional cases that each case produces [12] and the average time between the onset of a case and the onset of a subsequent case infected by that case [13]. The starting date for these projections was the date when each country had 25 confirmed cases reported in the World Health Organization Situation Reports (SITREPs); for countries who had not reached this number by March 24 th , the projections were made from March 25 th .

To perform the comparisons, we extracted data on the actual number of confirmed COVID-19 cases from the SITREPs of May 1 st [14] , May 29 th [15] and June 26 th [16] . The 2019 total population estimates were obtained from the United Nations World Population Prospects 2019 [17] .

Our study was exempt from Institutional Review Board approval since we used data already collected.

We first determined the percentage difference between the projected and actual number of COVID-19 cases by country on May 1 st , May 29 th and June 26 th . To do so, we calculated the difference between the actual and projected number of cases, divided it by the projected number of cases and converted the results to percentages. Further, we computed the projected and actual population proportion of COVID-19 infection by country on May 1 st , May 29 th and June 26 th ; we did it by dividing the total number of confirmed cases reported in the SITREPs by the total population count of each country indicated the United Nations World Population Prospects 2019. Thereafter, we compared the values of the projected and actual proportions of COVID-19 cases using the two-sided "N-1" Chi-squared test as recommended by Campbell [18] and Richardson [19] . Further, we performed Bonferroni correction to reduce the risk of type 1 error while performing multiple statistical tests [20] . The 95% confidence interval of this difference was estimated following the recommended method given by Altman et al. (19) . The value 0.05 was used as threshold of statistical significance for all the statistical tests performed. All the analyses were done in the statistical software R version 3.6.1.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted July 7, 2020. . https://doi.org/10.1101/2020.07.06.20147124 doi: medRxiv preprint

On May 1 st , May 29 th and June 26 th , respectively 40 (86.95%), 45 (97.82%) and 41 (89.13%) sub-Saharan African countries had a number of confirmed cases that was below the predicted number of 1000 cases for May 1 st , and 10000 for both May 29 th and June 26 th (table 1) .

the population proportions of confirmed Covid-19 cases were significantly lower (p-value <0.05) than the projected proportions of cases. (table 2) .

On May 1 st and May 29 th , all the confidence intervals of the difference between the projected and actual proportions of cases excluded 0. On June 26 th , 41 (75.92%) of the confidence intervals of the difference between these proportions excluded 0.

Across all these dates, South-Africa always exceeded the predicted number and population proportion of Covid-19 infections.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted July 7, 2020. 

This cross-sectional study comparing the projected and actual number of COVID-19 cases in the 46 sub-Saharan African countries found that on the dates specified, most countries did not meet the predictions in terms of number of COVID-19 cases. These results might be explained by the prompt response of African leaders, the distribution of COVID-19 high risk groups in sub-Saharan African countries, the disagreement between the statistical model parameters (or assumptions) and the reality, and the low testing and reporting capacity of sub-Saharan African countries.

From its first description [1] , the COVID-19 infection is characterized by a rapid spreading ability. Human hosts of the virus are major carriers that can favor its transmission if they move from place to place. Therefore, the reduction of imported cases was one of the first important preventive measures against the spread of the COVID-19 infection. African leaders were quick to react; on March 11 th 2020 the WHO declared the COVID-19 outbreak a global pandemic [2] , and as of March 15 th , most African countries had closed their borders until further notice [21, 22] . Thus, notwithstanding the harsh socioeconomic consequences, these results justify the positive impact of border closures towards limiting the spread of COVID-19. African public health authorities were also quick to implement and enforce hygiene sanitation measures. Finally, quarantine measures were imposed for several weeks in most countries [23, 24] .

The Centers for Disease Control and Prevention (CDC) identified persons at higher risk of COVID-19 infection as those of older age (above 65years) and/or those with severe medical conditions like chronic kidney disease (CKD), obesity, and immunocompromised states especially chronic stress/anxiety, type 2 diabetes mellitus (T2DM), and HIV/AIDS [25] .

The higher risk associated with older age is related to the weakening of the immune system that occurs with aging [26, 27] . Furthermore, children have been found to be at lower risk and to present milder symptoms of COVID-19 infection [28] [29] [30] . According to the United Nations Population Report [17] , in 2019 there were 45 526 000 people aged 65 and older in Africa, which represents 3.5% of the total population. This is quite low compared to the United States (US) and Europe where they represent 16% and 18.8 % of the total population respectively [17] . Furthermore, the African population aged 0 to 14 years makes 40.55% of the inhabitants of the continent. Thus, the natural composition of the sub-Saharan African population might explain part of its resilience to the COVID-19.

The higher risk associated with CKD, obesity, and immunocompromised states that are chronic stress/anxiety, T2DM and HIV/AIDS is also linked to the weakening of the immune system that characterizes these pathologies [31] [32] [33] [34] [35] . Despite a relatively higher burden of HIV/AIDS [36] and CKD [37] , the African population hosts lower proportions of individuals with chronic stress/anxiety [38] , obesity [39] and T2DM [40] as compared to the US and Europe. Obesity is a risk factor for several pathologies [39] which in turn are also favoring conditions for the development of the COVID-19 infection; such a synergistic effect between obesity and these pathologies might account for part of the higher COVID-19 burden seen in the US and Europe as compared to Africa.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted July 7, 2020. . https://doi.org/10.1101/2020.07.06.20147124 doi: medRxiv preprint

The difference observed between the projections and actual figures of the COVID-19 burden in Africa might also be explained by the fact that the epidemic parameters and /or assumptions used in the predictive model did not match with the reality on the ground. The epidemic parameters used considered that each case would produce an average of 2 additional cases [12] , and that the average time between the onset of a case and the onset of a subsequent case infected by that case (serial interval) would be 4.7 days [13]. These parameters presented several limitations. First, they were estimated in the Chinese population which substantially differs from the sub-Saharan African population in many aspects that can fundamentally impact the spread of the COVID-19 infection. Furthermore, Abbot et al., did acknowledge the scarcity of data at the time they built their model and they further admitted that their results would be significantly impacted by the availability of new data [12] . In the same line, Bi et al., did point out that their study had "numerous limitations", including the high risk of bias due to the multiplicity of data collection protocols, the fact that it was "impossible to identify every potential contact an individual has", and the fact that asymptomatic travelers were missed [13] . Finally, for such predictions to come true, the living conditions should remain stable in order to agree with the statistical model parameters. However, the COVID-19 pandemic is an ever-changing affection which prevent predictive models from describing it accurately.

It is important to note that the actual figures might be low because all cases might not have been reported. Because of the novelty of the COVID-19 infection, testing kits had to be made from scratch. Countries with technical and infrastructural abilities were the first to dispose of COVID-19 tests. The African continent suffers from a substantial lack of high technical capacities that can produce such quality tests in sufficient amounts; therefore, the continent entirely relies on other countries to scale-up its testing abilities. Because of the high global burden of the COVID-19 infection, countries with testing capacities did put their populations' interests first, leaving dependent countries like African ones on endless waiting lists. This, in addition to the overall weak health system (low financial capacity for purchasing tests; low man power capacity and communication infrastructures to ensure wide geographic coverage of testing teams or decentralization of testing laboratories) hampered the capacity of African countries to massively test their populations early, which might in part explains the low number of reported cases. For instance, the testing capacity measured by the number of tests performed per one million population is for instance 142325 in UK, compared to 672 in Nigeria [41] . Furthermore, African countries do not possess strong public health reporting systems, thus the number presented might not accurately describe current reality.

Finally, whether some specific individual or environmental factors have come into play by for instance reducing the virulence and spread of the virus, is still questioned. Research has suggested that vaccination with Bacillus Calmette-Guérin (BCG) can have protective effects against the COVID-19 infection [42] , and BCG vaccination is compulsory at birth in almost all sub-Saharan African countries as part of national enlarged vaccination programs. High temperatures and humidity have also been suggested to be protective against the COVID-19 infection [43] , and sub-Saharan African rests in high daily average temperatures and humidity percentages [44] . Research has also hinted at a protective effect of Vitamin-D against the COVID-19 infection [45] ; sub-Saharan African countries are under high sun exposure year-round which confer higher levels of Vitamin-D to their populations, compared to Western countries [46] . Studies have also described that African populations have specific . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted July 7, 2020. . https://doi.org/10.1101/2020.07.06.20147124 doi: medRxiv preprint genetic characteristics and adaptations that have modified their susceptibility to various infections [47], thus leading scientists to hypothesize about a potentially stronger natural resistance of sub-Saharan African populations to the COVID-19 infection Our study has three major limitations. First, there is a risk of ecologic fallacy since we used individual data to do our analyses but rather used population data to explain our findings. Second, statistical tests, p-values and confidence intervals carry their share of limitations and insufficiencies that need to be mentioned and accounted for. Finally, we used secondary data that were already collected, and therefore our study is exposed to the potential biases that might have arisen during the collection of such data.

Our study is the first to provide tangible evidence of the fact that sub-Saharan Africa truly defied the dire predictions of the spread of the COVID-19 infections. Our study can inform African stakeholders about their continent's unique advantageous characteristics that can form the ground for efficient and effective public health strategies against the COVID-19 pandemic, and beyond, against all the health hazards that deprive the continent from a much needed healthy population.

Contributors. DCFM and SPC conceived the project. DCFM drafted the manuscript, extracted the data, and performed the analyses. DCFM, AA, MYAN, MVK, SFM and SPC critically revised the manuscript for methodological and intellectual content. All authors approved the final version of this protocol.

Funding. The authors have not received any specific grant for this research.

The authors declare no conflicts of interests. The findings and conclusion in this manuscript will be those of the authors and will not necessarily represent the official position of the institutions to which they belong.

A pneumonia outbreak associated with a new coronavirus of probable bat origin | Nature

WHO Director-General's opening remarks at the media briefing on COVID-19 -11

Beijing orders 14-day quarantine for returnees -BBC News

Coronavirus Disease (COVID-19): Situation Report -166

Projection of early spread of COVID-19 in Africa as of 25

Africa will be next epicentre of coronavirus, says WHO. euronews

Coronavirus: Africa could be next epicentre, WHO warns -BBC News

COVID-19 Pandemic: Africa Defies Doomsday Prediction of Massive Deaths -Nigerian Democratic Report

Why does Africa continue to defy Covid-19 predictions? Nile Post. 2020

The problem with predicting coronavirus apocalypse in Africa

The transmissibility of novel Coronavirus in the early stages of the 2019-20 outbreak in Wuhan: Exploring initial point-source exposure sizes and durations using scenario analysis

13 Serial interval of novel coronavirus (2019-nCoV) infections | medRxiv

Published Online First: 1

Coronavirus Disease (COVID-19): Situation Report -130

COVID-19 Disease Response | Situation Report

Highlights | Multimedia Library -United Nations Department of Economic and Social Affairs

Chi-squared and Fisher-Irwin tests of two-by-two tables with small sample recommendations

The analysis of 2 × 2 contingency tables--yet again

When to use the Bonferroni correction

21 African nations close borders, cancel flights to contain coronavirus spread -Reuters

Why African nations close borders

African govts using a similar COVID-19 containment rule book. Africanews

What African Nations Are Teaching the West About Fighting the Coronavirus. The New Yorker

Centers for Disease Control and Prevention

Aging of the Immune System: How Much Can the Adaptive Immune System Adapt?

Causes, consequences, and reversal of immune system aging

COVID-19 in Children: An Epidemiology Study from China

Epidemiology of COVID-19 Among Children in China

Risks to children and young people during covid-19 pandemic

Psychological Stress and the Human Immune System: A Meta-Analytic Study of 30 Years of Inquiry

Aspects of Immune Dysfunction in End-stage Renal Disease

Impact of Obesity and Metabolic Syndrome on Immunity12

Type 2 Diabetes Mellitus and Altered Immune System Leading to Susceptibility to Pathogens, Especially Mycobacterium tuberculosis

The Immunocompromised Host

and forecasts to 2030, for 195 countries and territories: a systematic analysis for the Global Burden of Diseases, Injuries, and Risk Factors Study

Global, regional, and national burden of chronic kidney disease, 1990-2017: a systematic analysis for the Global Burden of Disease Study

Our World in Data Published Online First

Health Effects of Overweight and Obesity in 195 Countries over 25 Years

Epidemiology of Type 2 Diabetes -Global Burden of Disease and Forecasted Trends

Coronavirus Update (Live): 11,082,329 Cases and 526

PDF) BCG vaccination may be protective against Covid-19

The Effects of Temperature and Relative Humidity on the Viability of the SARS Coronavirus

Role of vitamin D in preventing of COVID-19 infection, progression and severity