Since the first COVID-19 case in Africa was recorded on Feb. 14, 2020 in Egypt, there have been more than 1.1 million reported confirmed cases and more than 26,000 deaths across the continent. This is a large total, but it is lower than the high figures many projected (1,2,3) at the onset of the pandemic. Despite having 17% of the global population, Africa has accounted for just 5% of global COVID-19 confirmed cases and 3% of global COVID-19 deaths.
In this Insight, we examine epidemiological information to better understand what we know about COVID-19 in Africa and what critical information gaps remain.
How much COVID-19 is there in Africa?
COVID-19 has spread in different ways at different times in different regions of the world. As shown in the chart below, Europe had a noticeable increase in per capita cases from February to April, South America from May through July and North America from June to mid-August. In contrast, Africa has had a consistently low per capita rate of confirmed cases since May 2020.
Despite a relatively late onset of COVID-19 cases, many Africa countries acted early to implement public health and social measures to prevent disease transmission. As of April 1, 2020 there were around 6,000 confirmed cases across all of Africa, with most countries having less than 100 cases. Despite this low level of detected cases, most African countries had stringent measures such as school closures, workplace closures and travel restrictions in place. This may have contributed to lower levels of disease introduction and ongoing transmission in the initial months of the pandemic.
Source: Our World in Data
Another reason for lower COVID-19 incidence could be more effective disease control to drive down transmission. Most countries in Africa routinely manage infectious disease outbreaks and have experience with disease control measures such as isolation, quarantine and contact tracing. Currently, many countries in Africa are doing comprehensive contact tracing for all cases.
Source: Our World in Data
To understand whether these contact tracing programs are effective, additional data is needed on their program performance including identifying and stopping chains of disease transmission.
Every country in Africa has reported confirmed COVID-19 cases. There is great variation in the number of cases by country, even after accounting for differences in population.
Source: Our World in Data
South Africa, which accounts for just over 4% of all people in Africa, accounts for over 51% of total confirmed COVID-19 cases (more than all other countries in Africa combined).
There are many reasons why COVID-19 transmission could vary in different locations. These include differences in the population (age, sex, race/ethnicity, socioeconomic status, behaviors, nutritional status, comorbidities, genetics, population density, household size), potential differences in the virus (strains and virulence), and differences in the environment (humidity, temperature). They also include differences in adherence to measures such as wearing masks, washing hands and physical distancing. Lastly, the effectiveness of response efforts could vary greatly by location. All of these factors likely contribute in some way to differences in the total number of COVID-19 infections in a particular area. A major reason for the observed variation in the number of confirmed cases is a difference in the ability to detect those who are infected. This is impacted by the capacity to perform timely diagnostic testing, and the presence of barriers (e.g. cost, stigma) to getting tested.
COVID-19 case counts are highly dependent on the level of COVID-19 testing, and areas with higher testing rates are more likely to find people who are infected. To understand the extent of COVID-19 spread, one must consider the rate and results of COVID-19 testing being done. The Our World in Data website routinely reports testing data by country, and most countries in Africa do not have frequently updated testing data.
Source: Our World in Data
Among countries that do have data, there are generally lower rates of testing (yellow shade) for countries in Africa compared to other countries.
Source: Our World in Data
This is expected, as the rate of testing (and health and essential services in general) correlates with resources in a country. Many African countries have lower per capita GDP and generally lower per capita testing rates compared to other countries around the world.
Source: Our World in Data
The level of testing required to have a more complete picture of disease spread depends on the amount of disease present. The more cases in an area, the more tests are required to try to find everyone infected. The figure below shows per capita testing rates and confirmed case rates. The farther down on the chart, the less adequate the testing (and the higher the test positivity). Many countries in Africa have suboptimal rates of testing given the number of cases they have. Of the countries highlighted, only Rwanda, Uganda and Togo have high rates of testing and low test positivity, suggesting that testing may be adequate to detect a higher proportion of cases.
Source: Our World in Data
The decision of who to test depends in part on the availability of tests. Locations that have an abundance of tests may choose to test anyone, including those who do not have symptoms. When sufficient testing is not available, prioritization is necessary, and testing may be restricted to people in priority groups such as those who are symptomatic, vulnerable or hospitalized. As expected, most countries in Africa target their testing to those with symptoms, as shown below.
Source: Our World in Data
Testing prioritized groups is appropriate when testing capacity is limited, but it also results in a limited understanding of spread of the virus, often missing people who have milder illness (e.g., younger people) or those who have no symptoms. These groups can also transmit disease, so limited testing can hamper efforts to control outbreaks by rapidly isolating cases to interrupt chains of transmission.
Notably, South Africa is one of a handful of African countries that offer open public testing, which implies that they have more robust testing capacity and will identify more infections occurring within the country, and relatively more cases than other countries in the region. Data on testing rates for countries in Africa with testing data available from August 10 to 17, 2020, show that countries with open testing policies have had higher total rates of testing.
Among the countries listed above, South Africa and Morocco also report “people tested” whereas all the other countries are reporting “samples tested” or “tests performed.” Reporting people tested rather than the number of tests performed likely results in relatively lower test estimates for these two countries in comparison to the other countries.
Looking at total tests by country, for countries with recent testing data available, South Africa makes up nearly half of the tests done in this group.
Test positivity rate, or the percentage of all tests (some use individuals) that are positive, is an important indicator to measure both whether testing is adequate to identify a large portion of true infections, and to understand the spread of the virus. Lower positivity rates are better, and WHO identifies a 5% positivity rate or less as one of the criteria to assess whether an epidemic is controlled. Looking at test positivity across Africa, we see that for most countries this information is not available.
Source: Our World in Data
For countries with data, we can see that most countries have positivity rates exceeding 5% (orange or red). Only three countries—Rwanda, Togo and Uganda—report test positivity rates less than 5%. South Africa, the country in Africa with the most cases by far, has a positivity rate of 15%, which is far greater than the optimal level and implies many undetected cases in the country.
Overall, these findings suggest that there is inadequate data on testing in Africa, and countries that offer tests more freely have performed more testing, which can help explain discrepancies in case counts. More frequent reporting of testing data from all counties is essential to understanding the extent of COVID-19 spread in Africa.
What can we learn from COVID-19 death data?
Death patterns over time also show a consistently low level of per capita deaths in Africa. This in in stark contrast with Europe which had high death rates in April, North America which has had high deaths rates since April, and South America which has had high death rates since May.
Over the past six months, we have learned that older age is a significant risk factor for severe illness and death from COVID-19. Estimates of the case fatality ratio (proportion of confirmed cases that die) and infection fatality ratio (proportion of total infections, detected or undetected, that die) are about 20 times greater for those 60 years or older compared to those less than 60. In the United States, eight out of 10 COVID-19 deaths have been in adults 65 and older. Across the world, there are large differences in the age structure of countries and regions. Consequently, we should expect that areas with older populations would have higher rates of deaths among COVID-19 cases.
In the above figure, we can see that Europe is the region with the greatest percentage of population 70 years or older and the highest documented COVID-19 case-fatality rate. In contrast, Africa is the youngest region, and as expected has had the lowest documented case-fatality rate.
Examining COVID-19 deaths in Africa by country, we see patterns similar to what we saw with cases, with deaths concentrated in a few countries.
South Africa, which accounts for 51% of total confirmed COVID-19 cases on the continent, accounts for 46% of deaths. Two out of every three deaths attributed to COVID-19 in Africa have occurred in South Africa and Egypt.
Deaths attributed to COVID-19 are just one indicator of the overall impact of the pandemic. The quality of death surveillance varies significantly by country. The limited testing described above also limits the number of deaths attributed to COVID-19, and every country underestimates the number of deaths from COVID-19 to varying degrees. To better understand the full impact of the pandemic, we should look at the total number of lives lost in excess of historically expected levels.
This metric not only captures deaths from COVID-19 (including diagnosed, reported cases and cases which are either undiagnosed or unreported), but also deaths from indirect impacts of the pandemic, such as delayed or inadequate access to routine health care. For most countries in Africa, this data is currently unavailable as civil registries often only produce annual statistics, and certificates for many deaths are never issued at all. Estimates of the excess deaths in South Africa reveal a huge increase since June—many more deaths than officially attributed to COVID-19. Overall, an estimated 36,587 excess deaths occurred from May 6 to August 11, 2020, far exceeding the 10,621 cumulative COVID-19 deaths reported as of that date. This likely represents a combination of unrecognized COVID-19 deaths and other deaths indirectly related to the pandemic (for example, from heart disease and tuberculosis). Among the estimated excess deaths, 24,621 (67%) were among those 60 and older.
Weekly reports of excess mortality such as those from South Africa should be produced by all countries, as these reports can inform policymakers as they navigate the pandemic. This data is not always available in a timely manner in many low-resource settings. Rapid mortality surveillance can fill this gap where existing civil registration and vital statistics (CRVS) systems are unable to meet the need.
How many people have COVID-19 in Africa?
The interpretation of cases in the context of uneven testing is challenging. To better understand the full extent of disease spread, one can use deaths to estimate the true number of infections in a country. The basic premise is that it is easier to ascertain COVID-19 deaths than it is to identify those who are infected. By taking the total number of recorded COVID-19 deaths and the number of infections that would typically lead to a death in a given population with a given age structure, one can estimate the total number of infections including asymptomatic and symptomatic cases.
The infection fatality rate (IFR) is the estimated rate of deaths in all who are infected (not just those that are diagnosed, which is represented by the case fatality ratio, or CFR). For COVID-19 the best global estimate of overall IFR is about 0.65%, or about 1 death for every 154 total infections, including those without symptoms. Countries with older populations can have IFRs exceeding 1%, and those with very young populations have IFRs approaching 0.1%. As with CFR, this number can vary greatly based on the age distribution of a population, treatments available and timeliness of treatment. Several studies have estimated IFR by age, and using these estimates and the known age distributions of countries, we can generate country-specific IFR estimates.
Using these estimated IFRs and the number of recorded COVID-19 deaths, we can get a rough estimate of true total number of “expected cases” in a population. For example, for South Africa we estimate there have been approximately 4.3 million total infections. In the population of 58 million people this is a cumulative infection rate of 7.5%. This is in line with the current estimate from other models such as the COVID-19 projections site which estimates 9% (95% CI 5-14%) of the population has been infected. Our estimate is likely underestimated to some degree since there are likely more COVID-19 deaths in South Africa than what has been reported (see above on excess death data). For Africa overall, we estimate there have been roughly 15 million total infections with a cumulative infection rate of about 1% to 1.5%.
We can then compare the expected total cases with the actual number of reported cases to generate an estimated “detection rate.” This rate represents the percentage of all estimated infections that have been detected. For this analysis we did not account for the lag between cases and deaths. Another limitation is that death reporting is incomplete (but likely better than case reporting).
As shown above, the rate of detection is highly variable, with Gabon detecting 43% of estimated infections and Chad detecting 2.3%. Overall, in Africa we estimate that only approximately 1 in 12 to 15 infections has been detected. We can also compare the documented CFR with the expected IFR to see which countries may be under-detecting cases. The closer CFR is to IFR, the greater the detection rate. The more CFR exceeds IFR, the larger the detection gap.
A better way of estimating the true number of infections may be to conduct a serosurvey to measure the prevalence of antibodies to SARS-CoV-2, the virus that causes COVID-19, in a population. Many serological surveys are currently being done worldwide and used to inform the current estimates of IFR. WHO has developed a general protocol for population-based age-stratified serological surveys. A challenge to conducting these studies is that diagnostic capacity is needed; in addition, the results can vary depending on the quality of the test used. In particular, many antibody tests return false positive results, and it appears that some antibody tests may return false negative results several months after infection. In general, the quality of antibody tests has not yet been well validated. In Africa, many serosurveys are being planned or in process, but few have been completed. Two (1,2) recent preprint systematic reviews of early serosurveys around the world did not include any results from Africa. One small serosurvey from Ethiopia examined 99 people in a laboratory waiting room from May 18-21, 2020 and found that three (3%) had evidence of prior SARS-CoV-2 infection. Serotracker, a new online resource that tracks serosurvey results, includes the results of only one study, from Kenya, in its data from Africa. In this study of 3,098 blood donors from April 30 to June 16, 2020, the population-weighted test-adjusted national seroprevalence estimate was 5.2% (95% CI 3.7-7.1%) This is several orders of magnitude higher than the estimated prevalence (0.5%) that is calculated based on known COVID-19 deaths in Kenya as of August and estimated IFR, and indicates that many COVID-19 deaths are not being captured (hence supporting the need for routine reporting of total excess deaths). Data from rigorously conducted serosurveys in representative sample populations, using validated antibody tests, are needed to triangulate our understanding of the true infection rate across Africa.
What other impacts from COVID-19 are important?
The ongoing pandemic has highlighted many false dichotomies, including open vs. closed societies (more of a phased approach), health vs. economy, and droplet vs. aerosol transmission. Many of these represent an oversimplification that obscures important details. With respect to the impact of COVID-19, a false dichotomy is that those who are infected either do well and recover (“it is just the flu”) or die. In reality, COVID-19 has proven to be a devastating disease among many survivors, during acute illness or after recovery when “long haulers” experience lingering symptoms affecting multiple organ systems and quality of life. A major concern is the impact of COVID-19 on health systems. COVID-19 hospitalization rates vary, but even if just 5% to 10% of confirmed cases require hospitalization, that represents 55,000 to 110,000 people in Africa and more than 25,000 to 50,000 in South Africa alone. Indeed, South Africa has had difficulty managing the stress of COVID-19 patients on the health care system.
In addition to complications from acute infection, some people with COVID-19 have had lingering, debilitating symptoms for months. In some studies of select groups of cases, most still had symptoms at two months, and many had heart abnormalities several months later. In one study, more than one in three outpatients with COVID-19 did not return to their usual state of health two to three weeks after testing. There are also adverse mental health impacts among those infected, those who have unintentionally infected others, those who are caring for the infected and the general population.
Together, these data suggest that the impacts of COVID-19 go far beyond just deaths, and even if a small percentage of those who are infected have long-term symptoms, this still represents a sizeable number of people who will experience ongoing health problems.
Other considerations in Africa
Many factors are likely to influence the observed patterns of COVID-19 infections and deaths in Africa. A younger age distribution, lower overall population density, warmer temperature, less urbanization, relatively early implementation of public health and social measures, better case investigation and contact tracing, and other factors common in Africa tend to favor less transmission and less severe disease. Conversely, larger households (see figure below), more multigenerational households, high rates of malnutrition, high rates of infectious diseases and other factors may lead to additional burden of illness from COVID-caused disruptions in health and social services in comparison to other regions. Within the continent, countries with hubs for international travel are more likely to have importation of cases in the absence of travel restrictions. The balance of these factors will drive trends in the true number of infections and the number of deaths.
As we previously noted, the COVID-19 pandemic has already had a dramatic, measurable effect on the capacity of health systems to support health. There is cause for concern that the indirect effects of the COVID-19 pandemic on health and health services may far outweigh the direct effects of the pandemic, particularly in Africa. This is likely to be particularly true in areas where the direct impact of COVID-19 is less severe and/or where there is a greater reliance on essential health services (e.g., where the risk of vaccine preventable diseases is greater, or where the burden of HIV, TB or malaria is high). The best way to minimize these impacts is to control COVID-19 transmission and protect health care workers and health care systems, so that essential health and social services can operate with minimal disruption.
Just over six months after the first known case of COVID-19 in Africa, many experts don’t agree on the reasons why there has not been an explosion in cases and deaths on the continent. In looking at the epidemiological information available, we can clearly say we do not have enough reliable data across the continent to have a comprehensive understanding of the impact of COVID-19. Low testing numbers make the low case counts unreliable. Very limited serosurvey data and estimates of total infections using death data both indicate that the actual number of infections is orders of magnitude greater than what has been reported. Without more widespread reporting of excess deaths, the mortality burden is not clear. African countries may have relatively better disease control performance given their experience dealing with outbreaks, but this is hard to ascertain based on available data. The importance of other factors in relation to COVID-19 is also largely unknown.
To address critical gaps in our knowledge of COVID-19 in Africa, we need: 1) Regularly updated testing data to understand limitations in case data; 2) Response data to understand if measures such as contact tracing are performing adequately; 3) Rigorously conducted serosurveys; to estimate the true prevalence of disease and inform IFR estimates; 4) Weekly excess mortality data to understand the overall impact of the pandemic on total deaths; and 5) Ongoing scientific research to understand how different factors relevant to Africa interact with COVID-19 transmission and severity of illness.