The irrelevance of containment as an Ethiopian COVID-19 response
By Daniel Hailu, July 21, 2020
(Ethiopia Insight) — With community transmission now widespread, Ethiopia needs to spend scarce resources on treatment
Maintaining containment as a COVID-19 public health response has become irrelevant in Ethiopia for two reasons. First, it is too late. Despite the lack of extensive testing, there is circumstantial evidence that the virus has already spread widely through community transmission—that is, transmission with no link to known confirmed cases. Second, the country simply does not have the institutional capacity to make containment effective.
The article’s primary source of information is participant observation that I had the opportunity to conduct at Millennium Covid-19 Care Center (MCCC) in Addis Ababa. I tested positive for the virus and was admitted to the center on 2 June, and stayed until I was discharged on 22 June.
During that period, I observed the services that MCCC provided and the dynamics of relationships among members of the MCCC community, including physicians, nurses, administrators, cleaners, food dispensers and patients. I spoke to doctors and nurses about their perceptions and management of the symptomatology of patients, as well as with several fellow patients about the path that brought them to the center—how they happened to be tested and notified about their status, and how their contacts were traced and managed.
On 19 June, the Ministry of Health changed some of its guidelines in the COVID-19 State of Emergency, particularly in relation to isolation of infected persons. While these changes are in line with my observations and recommendations, they are still within the framework of containment as a response, which this article argues against.
Low per capita testing
Testing, the bedrock of an effective containment response, is aimed at identifying infected persons so they can be isolated. International experience, such as in South Korea and India, has shown the critical value of ramped-up testing in containing the virus’s natural spread. The higher the percentage of a population tested, the more infected members are identified and isolated before they infect others.
Ethiopia, however, remains to have one of the lowest per capita testing rates in the world. It has only been able to increase testing capacity to about 5,000 per day in the twelve weeks since the first confirmed case on 13 March. In a country of more than 100 million people, this means only a small proportion of potential infections are being identified and isolated. Given these figures, it is clear that the current low testing capacity is insufficient for a containment response.
Widespread community transmission
International experience indicates that widespread community transmission makes containment virtually useless as a COVID-19 public health response. In the absence of extensive testing, there is no way of knowing the extent of community transmission—but there is abundant circumstantial evidence that the virus is already widespread in Ethiopia.
For one thing, the percentage of confirmed cases through community transmission has risen rapidly. In the eighth week (8 May) after the first confirmed cases, community transmission accounted for about 21 percent of the total. However, according to official data that proportion has risen to over 65 percent over the past eight weeks. If Ethiopia was able to increase testing, there is a high likelihood that several thousands, if not millions, of additional infections would be found and would be attributable to community transmission.
Another indicator is the number of cases identified through random testing. Several fellow patients at the MCCC told me they had been found positive after dropping by a testing center out of curiosity about their status. Others went to a hospital to visit sick relatives, saw people testing and gave their sample. Yet, others, like myself, went for treatment of a minor ailment such as gastritis and their doctors referred them to test for COVID-19, apparently to minimize their own risk of infection from interaction with a potentially infected client. All such cases reported no contact with confirmed infections. This suggests that at least in Addis Ababa the chance of any person testing positive is substantial.
The post-humous confirmation of several cases also suggests widespread community transmission. Of the 61 COVID-19 related deaths in Ethiopia between 19 May–19 June, 28 were confirmed through post-mortem forensic examinations. That close to 50 percent of total confirmed deaths occurred in communities during recent weeks implies the likely widespread of the virus.
The rapid growth in the numbers of infected cases in rural areas of Oromia, Afar, SNNP and Amhara in the early weeks of the pandemic is also suspicious. This was unexpected given that testing in Ethiopia has been mostly aimed at city residents, while in much of rural Africa the virus has spread relatively slowly.
If, as it seems, community transmission is widespread, the Eurocentric model of COVID-19 morbidity and mortality would suggest that Ethiopia is in for an increase in COVID-related hospitalizations and deaths However, as I noted in a previous article, based on the latest findings, the virus does not appear to be as severe in Ethiopia as in other hard-hit countries, which means the overwhelming majority of unconfirmed cases may be asymptomatic to moderately symptomatic, and may go unnoticed in the statistics.
This hypothesis is supported by the ratio of clients admitted to MCCC per various levels of symptoms, which was recently calculated in response to requests from higher authorities. The “Symptomatology” survey found that of approximately 800 clients in the center, 93 percent were asymptomatic while over six percent had mild to moderate symptoms. There were only four cases that needed oxygen support, far less than one percent. And there were no deaths during my entire stay at the center.
To be fair, there is an admission bias at MCCC against severe cases. At the time the survey was done, the center’s Intensive Care Unit (ICU) facilities were not yet in operation, so it was not accepting seriously ill clients. However, the bias may not be that significant. According to the Health Ministry’s daily situation update, there were fewer than 40 severe cases in all COVID-19 treatment centers in the country when the MCCC report was issued. Moreover, sources reported that the extremely limited COVID-19 dedicated ICUs at various treatment facilities have not yet reached capacity.
Isolating infected persons
Ethiopia puts confirmed cases in isolation centers until they recover from the virus as a strong element of its COVID-19 containment response. In addition, an isolation center may provide treatment to infected clients as needed. The isolation centers in Ethiopia have varying accommodation and treatment capacity. The MCCC, where I was quarantined, is the largest such facility in the country, with 1,300 beds.
What follows is a review of the center’s performance, based on my experience. Several metrics can be used to measure a center’s effectiveness, but I will focus only on four: management, testing, maintenance of case records, and nutrition.
Management: It seems that the MCCC is not being overseen by professionals trained and experienced in the management of similar entities. Upon admission, new clients are not told what services they might expect to receive, or how long they might be required to stay. Nor are they informed of the norms of behavior expected of them while at the center. Consequently, clients routinely complain about various aspects of the center, while managers struggle to explain and fill gaps as they go along. It is clear that potential challenges of managing the isolation centers were not anticipated and solutions were not provided in a standard operational manual; or, if such a manual exists, it is not being followed.
Testing and retesting: Ministry of Health guidelines adopted from the WHO state that each client in an isolation center shall be tested after three consecutive days without any COVID-19 symptom or seven days after the test that first confirmed the infection. In either case, a person who tests positive will not be discharged until two consecutive subsequent tests are negative.
This guideline is not carefully followed, however, at least at the MCCC. For example, I was tested five days after I showed no symptoms and nine days after my first positive test. It took five days to learn my test result, and that of several others tested the same day. This delay in analyzing samples in an environment that is highly infested with the virus is putting recovered clients at a potential risk of re-infection. In fact, several patients who tested negative on the first test after being admitted had positive result the next time. Moreover, the delay wastes resources that could better be used on new clients, and significantly increases the unit cost of treatment and accommodation.
The problem is further compounded by lack of order in scheduling tests and retests and announcing results. Many patients complained about having to spend at least seven days without any test while other patients who tested positive were being retested a second time. Still others gave samples before being told the result of a previous test.
Implementation of the Case Record Form: The Case Record Form (CRF) recommended by the WHO, and the associated electronic data capture system serve two purposes: They document each client’s medical history to enable systematic monitoring of the client’s condition, and they help build a clinical data base on COVID-19 symptomatology in the Ethiopian context to inform a locally relevant public health response.
At the MCCC, however, completion of the CRF serves mainly the first purpose. So, while the medical histories of moderate and severe cases may be documented, it is considered a luxury to document asymptomatic or mildly symptomatic cases, so they are routinely overlooked. For example, follow up of my case begun to dwindle two days after I was admitted when my minor cough disappeared. In the subsequent days, they stopped measuring my vital signs even once every 24 hours. Later, after three days without any follow up on even my vital signs, I complained. In response, the nurse said: “Please do not bother us! You are all fine. Our focus is on problematic cases”. Even when other nurses measured my vital signs to appease me, they did not record the results. I observed other asymptomatic and mildly symptomatic cases being treated the same.
Perhaps more disconcerting is that not all essential items in the form are completed. For example, a ‘Preparedness and Response Daily Situational Report”, jointly issued by the Ministry of Health and the Ethiopian Public Health Institute as at 29 April, reported that there was no data available on the severity of 26 percent of confirmed cases. I also overheard a supervisor of the physicians admonishing them about incomplete forms.
Finally, available information on each case is not being entered into the associated electronic data capturing system, making it increasingly difficult to collect and analyze the aggregate clinical data. For example, according to the staff members, the center had to manually review more than 800 CRFs to compile data in response to a recent request by higher authorities for a report on the number of clients broken out by the degree of their symptoms (asymptomatic, mildly, moderately and severely symptomatic). Such manual analysis inevitably suffers from inaccuracies and lack of details, making it next to impossible to generate the rigorous clinical data needed to make intelligent policy decisions.
Meeting nutritional needs and preference: Isolation centers need to ensure that the nutritional needs and preferences of each client are met. This is particularly critical for moderately symptomatic cases with comorbidities such as diabetes, hypertension and gastro-intestinal disorders. Moreover, there are several clients with religious food preferences.
Ethiopia’s case-management protocol in health care facilities prescribes that clients should be provided with “food, water and other basic needs timely’. However, meals were often served late. For example, breakfast was served not earlier than 9am and sometimes as late as 10am. Moreover, the protocol does not anticipate and address nutritional challenges that may arise. In the case of MCCC, provision of food is outsourced to hotels. We were told that the center had assigned a nutritionist to guide providers in tailoring food to the needs of various categories of clients. However, nearly three weeks after opening, MCCC was unable to meet the nutritional preferences of clients. For example, packed food labeled “for diabetics” contained food items rich in carbohydrate such as white bread, or products that contain salt. This is unacceptable for clients with diabetes and hypertension.
Physicians at the center said the challenge of managing comorbidities has more to do with nutrition than the virus. In terms of managing diabetics, for example, they complained that they had to administer insulin to clients whose sugar level could better be managed with appropriate nutrition. Moreover, attempting to manage comorbidities with medicine rather than nutritional adjustment further exacerbates the heavy financial burden that COVID-19 is imposing on Ethiopia.
According to WHO recommendations that Ethiopia has adopted, tracing the contacts of infected persons immediately after confirmation with the aim of testing them and isolating carriers of the virus is an effective containment response. An interim WHO guidance on contact tracing prescribes administering detailed interview with a confirmed case to identify contacts at various distance, duration and settings. Each of these reported contacts are then to be traced and tested as expediently as possible. However, contact tracing even in developed countries such as the United States involves a number of challenges.
Ethiopia might have been able to implement rigorous contact tracing in the early weeks when the number of confirmed cases was small. Anecdotal evidence from clients at the center, however, suggests that contact tracing became increasingly haphazard as the number of confirmed cases increased. For example, none of the clients consulted reported an interview by implementors to build an exhaustive list of their contacts. In some cases, implementors simply targeted family members and colleagues of confirmed cases. In other cases, no attempt has been made to even trace and test family members or colleagues. This means that only a small percentage of contacts of confirmed cases are traced, tested and isolated when positive, apparently making investment in contact tracing wasteful.
Moreover, standard procedures did not seem to be followed by implementors in the way contact tracing was done. For example, several clients reported that implementors came to their home with ambulance, took them and returned to their home the next day to simply measure the temperature of family members and tell them to self-isolate in their home without requiring them to test. In other cases, implementors visited the homes and offices of a confirmed case. In a variation of this, implementers took samples for each contact while the entire neighborhood of the confirmed case was put under police custody and his workplace closed until results were announced. In another, implementors simply counseled and encouraged contacts to visit a testing center and give samples. In other cases, contacts tested negative for the virus were asked to retest after 14 days.
The ethnographic evidence presented here shows that much of Ethiopia’s COVID-19 containment response has been irrelevant because it does not take into account the stage of the virus’ transmission. It also fails to consider the limited implementation capacity available at the community level, which has limited their sustainability, effectiveness and efficiency. Hence, an urgent need for a revised national COVID-19 strategy that is simultaneously based on local clinical data and less complex for local capacity to implement.
I recommend that such a strategy is composed of four pillars:
- Focus on building capacity for treatment
Available local clinical data such as the MCCC clients’ symptomatology report cited earlier seem to indicate that less than one percent of COVID-19 infections in Ethiopia would require hospitalization. Although dedicated COVID-19 treatment facilities have absorbed the small number of severe cases so far, one percent of all infections may become too large for existing facilities to handle in a country of over 100 million. Hence, it may be wiser to build capacity for treating severe cases—a capacity that would significantly strengthen the health system because it would continue in place even after COVID-19 is under control.
On the other hand, the enormous cost of maintaining isolation and quarantine centers has largely been wasted because the country’s testing capacity is too low to identify more than a small minority of infections. Hence, these facilities should be immediately closed down. Instead, the over 99 percent of COVID-19 cases, that are asymptomatic to moderately symptomatic, should be advised to self-isolate until the virus runs its course. To this end, a well-conceived Social and Behavioral Change Communication (SBCC) action plan is needed.
- Implement a SBCC action plan
In the interest of reducing the heavy burden on governments, it is critically important to devolve responsibility for prevention and management of less severe cases to local communities, households and individuals. A recent high-level conference I attended identified several shortcomings in current public education initiatives. Among them were the heavy reliance on radio and TV, which excludes low-income groups, as well as weak messaging and a wide disparity in effectiveness of campaigns by different institutions.
I recommend devising a new COVID-19 SBCC action plan that addresses these shortcomings. Essentially, the plan should segment the general population according to socio-economic categories and creatively define key messages and communications media appropriate to each category. It should, among other things, promote facemasks and personal hygiene as part of a prevention plan, as well as educating communities on how to determine severe cases and access treatment for them while managing non-severe infections at the local level. Various stakeholders may specialize in crafting appropriate messages and managing media targeted at specific population groups. The plan envisions creating a strong oversight body to coordinate the work of various stakeholders and engage them in improving behavior modification strategies.
- Repurpose testing
As noted earlier, the main objective of testing at the stage of containment is to identify infected persons and isolate them. Since the virus appears to have already spread widely, and given the low testing capacity, it would be futile to continue using testing as a means of identifying infected persons.
Instead, the Health Ministry should repurpose the current capacity of about 5,000 tests per day to preventing infections at health facilities for two reasons: First, granting the above recommendations that the health system focus on severe cases, there is a need to test suspected severe cases in real time before admitting them as COVID-19 patients. Secondly, if infected patients visiting health facilities are not tested and isolated, they may infect other patients, increasing the risk of fatalities due to the virus and reducing health-seeking behavior by the general public due to fear of infection. Moreover, it can also increase the chance of exposure by healthcare providers, potentially depleting an already-constrained number of healthcare workers.
- Build capacity for systematic surveillance
In the absence of ramped up testing, there is a need to build capacity for systematic surveillance to generate reliable data that can inform adjustment to Ethiopia’s COVID-19 public health, social, economic and political responses. To this end, we should welcome the government plan to administer antibody surveys, which should be conducted at regular intervals to a representative sample population at systematically selected data-collection sites. Breaking down the data by geographic and population characteristics would offer a rich body of local information on the virus’s past and current level of penetration. This would help policymakers adjust Ethiopia’s public health and economic responses.