It is hard to explain Ethiopia’s low infection rate if it’s assumed COVID-19’s impact is universal. Alternative theories are needed.
On 8 May, Ethiopia ended the eighth week since the first COVID-19 case was confirmed on 13 March with a total of 194 confirmed infections.
In countries with the highest rate of infections such as Italy and Spain, the numbers at that epidemiological timepoint were over 65,000. In African countries with the most infections, such as South Africa, Morocco, Algeria, Cameron and Egypt, the totals were between 1,794 to 5,951.
In East Africa, so far the least affected region on the continent, Ethiopia has one of the lowest per capita confirmed COVID-19 infections. Ten weeks after the first case, Ethiopia’s epidemiological graph continues to show a low increase, offering one of the best demonstrations for a recent WHO report that the virus is spreading more slowly in Africa:
Ethiopia’s low rate is puzzling considering the fact that WHO had listed it as among the most at-risk African nations because of its relatively high air connectivity and urban population density. Moreover, a continued spike in confirmed cases in the coming days and weeks will not solve this puzzle, unless it is accompanied by a corresponding increase in hospitalizations and deaths, which have stagnated at less than ten for a few weeks now.
Why, then, are the numbers of COVID-19 confirmed cases, hospitalizations, and deaths in Ethiopia so low?
If assumptions are based on COVID-19 being as infectious and as fatal in Ethiopia as it has been elsewhere, it is hard to explain. Instead, alternative explanations from recent scientific papers suggest that several factors may be compromising the infectiousness and/or fatality of COVID-19 in contexts like Ethiopia.
One explanation forwarded in Ethiopian discussions based on the standard assumption is that exponential increase in cases, hospitalizations, and death have been delayed because Ethiopia closed its borders and introduced mandatory quarantine just in time. The suggestion is this successfully prevented importation of sufficient index cases that could have triggered exponential increase before the eight-week landmark.
There is no way of knowing for sure how many index cases were imported before the introduction of mandatory quarantine, but conservative estimates based on the available data reveal sufficient imported infections to trigger exponential increase in community cases earlier than eight weeks.
To demonstrate, we know from sources that Ethiopian Airlines (ET) landed 1.3 million passengers at Bole International in February and March when the virus was ravaging Asia, Europe and North America. If we assume 20 percent of those passengers had Ethiopia as their destination and the remaining 80 percent were transit passengers, the average number of weekly ET arrivals during those two critical months would be 31,769. This excludes passengers on other carriers. Nor does this include land entry.
To estimate the proportion of infected passengers, we may use the total number of people arriving between 23 March—when mandatory quarantine was introduced—and 9 April, which is 4,327, an average of 818 per week; plus index cases confirmed during the first four weeks after mandatory quarantine, which was 80, or 20 a week. This means that during the first month after mandatory quarantine, there were an average of 20 imported cases every week from 818 passengers. An extrapolation of that to our earlier conservative estimate of 31,769 arriving weekly gives us 776 cases imported weekly during the six weeks between 15 February and 21 March. If we assume only 25 percent of that to be accurate, we can safely conclude that 194 case were imported a week in the six weeks before quarantine was introduced.
Let us now consider if an estimated 194 imported cases per week for six weeks is large enough to generate exponential increase in the course of the 13 weeks from 15 February to 8 May.
Experts tell us that depending on various factors, each person infected by the virus can infect two to four others, depending on social distancing measures and other factors, before recovering or dying within two weeks. Let us adopt a transmission rate of 2 (according to sources, Ethiopia applied an infection rate of 2.8 in its projection modeling). A cursory calculation applying that rate and the estimated 194 imported cases per week results in a total of at least 64,731 by 8 May.
This means there were sufficient index cases to trigger exponential increase imported before Ethiopia controlled its borders.
Testing for COVID-19
The low number of confirmed cases compared to this projection of infections on the thirteenth week could be attributed to Ethiopia’s low testing capacity because the number of confirmed cases is not a good indicator of how widespread the virus is. Indeed, Ethiopia has among the world’s lowest per-capita testing for COVID-19. However, Ethiopia has ramped up testing over past weeks. In early March, Ethiopia was sending samples to be tested in South Africa. By 21 April, Ethiopia had 27 testing laboratories, with a capacity to daily conduct about 6,000 tests. The average actual testing is now at around 3,500 a day.
Moreover, because it is far from overwhelmed by COVID-19, Ethiopia’s health system has had time to undertake more proactive testing particularly during the weeks prior to and after 8 May, and targeted vulnerable populations. However, this increase in proactive testing has not yielded a corresponding increase in cases. In fact, Ethiopia has among the lowest number of confirmed cases per 1,000 tests in the world.
One may yet insist that a substantial number of infections escaped the radar of increased testing. If that was the case, and COVID-19 was as universally fatal as it is assumed to be, Ethiopia would have experienced an increase in COVID-19-related hospitalization and death by now.
According to global norms that based on the universal infectiousness and fatality of COVID-19, 20 percent of all infections require hospitalization, five percent Intensive Care Unit (ICU) admission and one to two percent die. Not to mention fatality rates in hardest hit countries in Europe, Asia and North America, African countries such as Algeria, Egypt, Morocco, South Africa and Cameron that are experiencing significant fatalities reported from 58 to 348 deaths in the fifth week after the first confirmed death.
Building on our earlier very conservative estimates, Ethiopia should have experienced at least 647 deaths during the same period. The number of confirmed deaths has, however, been less than half a dozen. Moreover, although, according to the same estimate, COVID-19 would have required 12,946 hospitalizations and 3,237 ICU admissions by 8 May; yet less than a dozen cases have required intensive care since the first case on 13 March.
Still, there is the argument that low community awareness and poor health seeking in Ethiopia, means most cases have not been reported and people may be dying in communities. This, however, underestimates Ethiopia’s disease surveillance and response system, weak though it may be. Take, for example, the most recent most recent Yellow Fever outbreak. On 3 March, the Ethiopian Public Health Institute (EPHI) reported three suspected cases in rural Gurage Zone in Southern Nations region. If the system could identify an outbreak of a relatively obscure disease and of such a small size outside Addis Ababa, it should be able to identify COVID-19 cases and fatalities, given that the symptoms have been widely publicized and they would most probably be concentrated in Addis Ababa.
Moreover, Ethiopia has instituted additional COVID-specific measures to detect symptomatic cases and fatalities that include publicizing reporting hotlines for each region and mobilizing the media to increase awareness. A report at a conference noted that by 29 April, a total of 7,638 calls were received and responded to via the COVID-19 hotlines.
In addition, extensive surveillance campaigns have been undertaken that involved a three-stage process: door-to-door pre-screening of body temperature of millions of community members in high risk areas in various regions; further screening by rapid response teams; and then actual testing of suspected cases. Contrary to expectations, these efforts, however, did not yield increases in the number of daily cases.
One may, as a last resort, attribute the low volume of daily confirmed cases to the containment interventions (social distancing, public hygiene, community awareness) Ethiopia has implemented. The government is indeed taking a lot of measures, including: educating the public about the virus, its impacts and precautions to prevent spread; implementing social distancing measure such as closing schools; limiting capacity in public transport and making work from home arrangement for public servants. It has also enforced a State of Emergency to facilitate adherence to restrictions. Yet there is a consensus that compliance with them has been far from sufficient to curtail transmission of a highly infectious virus that could even find its way into well-protected zones like the White House, 10 Downing Street, and the Kremlin.
What else, then, can explain the low rates of confirmed cases, hospitalizations and deaths in Ethiopia ten and seven weeks after first confirmed infection and deaths respectively? Could it be because the rate of transmission, i.e. the virus’s infectiousness, is much lower? And/or could it be because almost all infected persons have been asymptomatic or mildly symptomatic to be missed by a surveillance system that focused on symptomatic cases and at-risk populations?
In other words, could it be because COVID-19 is not as infectious and/or as fatal in the Ethiopian context as universally believed? Hence, there is arguably a need for a different set of explanations that are not confined to this assumption.
There is some recent research on five factors, perhaps among others that are yet to be examined, that may compromise the infectiousness and fatality of the virus among specific populations and contexts. This may explain the relatively weak impact of COVID-19 in Ethiopia. These are climate, altitude, a young population, the BCG vaccination, and genetics.
A 24 April article by an Ethiopian clinician and former health minister, Yifru Berhan, entitled “Will Africa be devastated by COVID-19 as many predicted?” reviewed how the tropics, in which Ethiopia lies, was inhospitable to the family of coronaviruses, such as MARS and SARS, that COVID-19 belongs to, and predicted that the same was true of COVID-19—although it has fared relatively better than those other coronaviruses in the region.
This conclusion is supported by a study entitled “Spread of SARS-CoV-2 Coronavirus Likely to Be Constrained by Climate”, which compared the pattern of COVID-19 transmission in various climatic conditions and established correlations between low rate of COVID-19 infections and the tropical climate. Later a systematic review of 517 empirical studies on the effects of temperature and humidity concluded that warm and wet climates seem to reduce the spread of COVID-19. The actual mechanisms by which tropical weather may be suppressing transmission of COVID-19 should be a subject of further investigation.
One study among several others on the SARS coronavirus found that it was stable at low temperature and low humidity environment, facilitating its transmission in community in subtropical area, such as Hong Kong, during the spring and in air-conditioned environments. It also found that high temperature and high relative humidity in the tropics constrained the virus’s stability and, hence, transmission. Perhaps the same mechanisms may be at work in reducing the infectiousness of COVID-19 within the tropics, which have suffered relatively little from the novel coronavirus so far with only four out of 92 tropical countries hard hit.
Other studies found sustained exposure to sunlight in the tropics providing Vitamin D that increased immunity. One such study found a significant positive correlation between sunlight exposure and COVID-19 patient recovery. A review of scientific evidence of associations between Vitamin D and COVID-19 found that populations living in northern latitudes (such as the UK, Ireland, Northern Europe, Canada and the northern parts of the USA, northern India and China), had poor Vitamin D levels, especially in winter or if confined indoors, which made them susceptible to morbidity and mortality due to the virus.
A paper published on the U.S.’s National Center for Biotechnology Information website entitled “Does the pathogenesis of SAR-CoV-2 virus decrease at high-altitude?” reported on the potential association between COVID-19 transmission and altitude. They compared epidemiologic data of COVID-19 in high and low altitude regions to test the hypothesis that people who live at more than 2,500 meters above sea-level were less susceptible to developing severe effects from COVID-19. They suggested that physiological adaptations that counterbalance reduced level of oxygen at high altitudes may protect people from severe impact. They also suggested altitude may compromise the virus’s life span, reducing its infectiousness.
The relationship between age and COVID-19 fatality is perhaps the most widely investigated theme, and on which the strongest consensus is emerging. A systematic review of 45 studies entitled “Systematic review of COVID‐19 in children shows milder cases and a better prognosis than adults” reported that children accounted for one percent to five percent of diagnosed COVID‐19 cases, often had milder disease than adults, and deaths had been extremely rare.
Another study published by Science Direct called “Clinical features of COVID-19 in elderly patients: A comparison with young and middle-aged patients” concluded that the mortality of elderly patients with COVID-19 was higher than that of young and middle-aged patients and that elderly patients with COVID-19 were more likely to progress to severe disease. These studies suggest that in countries with younger populations such as Ethiopia—about 70 percent of the population is under 30—COVID-19 infections may tend to be asymptomatic or mildly symptomatic requiring much less hospitalization.
Presence of a policy Bacillus Calmette-Guerin (BCG) vaccination (primary used against tuberculosis) may also compromise the fatality of the virus. Last month the Economic Times published an article entitled “US scientists link BCG vaccination with fewer coronavirus cases, Indian scientists hopeful but cautious” which reported on a study which found that countries without universal policies of BCG vaccination, such as Italy, the Netherlands, and the U.S., had been more severely affected compared to countries with universal and long-standing BCG policies such as many developing countries including Ethiopia.
The study hypothesized that BCG vaccine may offer a degree of protection against COVID-19 infections or/and severity of symptoms. Another study from Fujita Health University called “Association of BCG vaccination policy with prevalence and mortality of COVID-19” found ‘quite substantial’ effect of BCG in providing protection against COVID-19 and noted that the total number of COVID-19 related deaths in the U.S as of 29 March would have been 20 percent of the actual figure, if the U.S. had instituted the mandatory BCG vaccination several decades earlier.
Genetics is another factor that may be responsible for varying severity and fatality of COVID-19. Scientist have long-established difference among populations in their susceptibility to infections and diseases. In October 2016, the journal Cell published two studies that traced differences in large part to genetics directing the way the immune systems of people with European and African ancestry work. Future research may show if genetic difference among these populations resulted in varying responses to COVID-19.
To conclude, the modeling used by the Ministry of Health wildly exaggerated the increase in infections in Ethiopia. Since then, the Ministry has not made public any revision of its projections. This belies the fact that a model may not offer useful policy guidance as long as the foundational assumption regarding infectiousness and fatality are imported. Indeed, making policy on the basis of such models may unnecessarily extend the heavy social, economic and political burden the nation has been sustaining thus far. For example, social distancing measures informed by imported clinical assumptions are debilitating the informal economy which sustains the majority of the urban population. They have also required postponement of a much-anticipated general election, subjecting the nation to the risk of constitutional and political crisis.
To correct for these critical modelling errors, there needs to be systematic building of local data that can eventually inform model-building whose projections can adjust Ethiopia’s policy response to the reality on the ground. To this end, Ethiopia should rigorously implement WHO recommendation on gathering and storing coronavirus clinical data. Although the Ethiopian national guideline also prescribes the WHO recommendations, quarantine centers do not seem to scrupulously implement the WHO Case Record Form and store it in the associated electronic data capture system. A dataset that would build out of thorough application of these instruments could over time show patterns in COVID-19 symptomatology in the Ethiopian context to inform local projections on related morbidity and mortality.
However, a body of data on the extent of current and past exposure to the virus can also be generated quickly from a combined antibody and diagnostic test survey administered to the same representative sample population in large cities in Ethiopia—with Addis Ababa as a pilot study area given its relatively high burden of confirmed infections. Assume a diagnostic test would find 20 percent of the sample population positive for the virus and the remining 80 percent negative. Antibody tests on that same population would show the extent to which the 80 percent had previous exposure to the virus.
In other words, reading results of the two tests together can show the virus’s current and past spread. Disaggregating the data by geographic and population characteristics would offer a rich body of information that could enable the government to adjust its public health and economic responses. Hence, investing in this survey immediately is highly recommended because it can put policy-making on more secure scientific ground than any model can ever hope to.
This is the author’s viewpoint. However, Ethiopia Insight will correct clear factual errors.
Editor: William Davison
Main photo: A busy Merkato bus station despite COVID-19 in the days before Easter; 18 April 2020; Maria Gerth-Niculescu
Query or correction? Email us
Published under Creative Commons Attribution-NonCommercial 4.0 International licence. Cite Ethiopia Insight and link to this page if republished.