Study of the COVID-19 epidemic in Belarus (part 1, November 2020)
The article is the English version of the study conducted in November 2020.
The entire world is currently experiencing the second wave of coronavirus and Belarus is not an exception. The official data from the Ministry of Health tried to convince us that the situation is not more serious than in the spring. However, everyone’s personal experience, as well as the absence of serious restrictive measures, and the data on the new cases from other countries cast doubt on the official figures in Belarus.
Spoiler alert! According to our results, the incidence rate is understated severalfold.
The Ministry of Health of Belarus provides statistics only on the total number of cases and deaths. That is why we defined several goals for our study:
- Assess the scale of the epidemic — obtain the most realistic figures on the total cases for the country as a whole and for each region.
- Assess the effectiveness of using personal protective equipment (PPE) and provide recommendations.
- Identify the attitude of the population to the epidemic, the measures to fight the infection, and the provision of medical help.
- Clarify risk groups, the severity, and the duration of the illness for different population groups given their symptoms.
- Identify outbreak areas.
- Assess the scale of the epidemic, obtain the most realistic figures on the total cases for the country as a whole and each of the regions.
The survey was conducted from 6 to 18 November 2020.
The survey was conducted via a Viber-bot (Viber is the most popular messenger in Belarus). Each respondent received a link to a Google Form with a unique identifier.
The survey was completely anonymous. No personal data including phone numbers were collected and registering with Google was also not required.
New respondents were added to the already existing database of subscribers to the People’s Poll Viber bot through advertisements in various public groups of the popular messengers and news portals’ social media pages. A total of 31,273 forms were collected, which yielded 25,500 responses after the data was cleaned up.
The following responses were excluded from the final assessment:
- citizens of countries other than Belarus;
- people who do not reside in Belarus;
- people under the age of 18, as there were too few of them to be grouped for analysis;
- people with incomplete data on gender, age, region, type and place of residence and education;
- people with incorrect or duplicate identifiers (the results used the form with the latest timestamp).
Processing of the results
25,500 responses were processed to obtain the estimates of the parameters of interest.
The fact that those who have had the disease are more inclined to take part in the survey may have distorted the results. To avoid this, two random samples of 10,000 respondents each were drawn from the existing People’s Poll audience, and the share of those who had the disease was estimated based on their answers. Cases, where the respondent received but did not fill out the form, were also taken into account. Estimates based on these random samples were checked against each other to validate the results.
The resulting sample also has socio-demographic biases that are common for the Internet audience: the largest subgroup of those who took the survey is comprised of city dwellers with higher education under the age of 40. However, the MRP (Multilevel Regression with Poststratification) method we used allows us to even out these biases utilizing statistical modeling and reweighting the results to get estimates for the country’s entire adult population taking into account the number of people in a particular social group within the population. This method provides not only the most likely estimate of any indicator, for example, the share of cases in the population, but also a credible interval in which this value is located. We are providing the 95% credible intervals below. This means that we can say that the estimated parameter lies within its limits with a posterior probability of 95%; in other words, the posterior probability of an indicator being outside its limits does not exceed 5%.
We have estimated that as of 18 November, 8.4% of the adult population of Belarus is the most probable value for the share of adult Belarusians having had the disease with a confirmed diagnosis (95% credible interval — between 5.7% and 12.0%). This corresponds to around 635,000 people (95% credible interval — between 434,373 and 909,890). If we additionally take into account those who believe that they had been infected but did not receive an officially confirmed diagnosis (it is known that not everyone is allowed to take the test and that not everyone seeks medical help), then the most likely value is 1.9 million people (95% credible interval — between 1.6 and 2.3 million).
It should be noted that these figures are likely to be even higher today since the survey was conducted before 18 November, making it quite probable that over 700,000 people have had the disease with a confirmed diagnosis as of 4 December.
Interestingly, men get sick more often. This may be related to the fact that men are more careless about PPE as our estimates show.
In the figure above, we see a pattern similar to the statistics for other countries: the second wave of coronavirus is about three times worse than the first one with most cases recorded in November. According to our data in November (by 18 November), the average number of diagnoses per day exceeded 6,300, while the Ministry of Health reported a maximum of 1,621. If we also count those who believe that they had been sick without an official diagnosis then the average number of cases per day is close to 15,000 people.
The numbers we obtained are very similar to Sweden’s morbidity statistics. Sweden is comparable to Belarus in terms of the climate and population as well as the anti-virus measures introduced by the government. For example, the November figures for the daily increase in cases in Sweden are quite similar to our assessment of the situation in Belarus.
To assess the epidemiological situation in Europe the European Center for Disease Prevention and Control (ECDPC) publishes a map based on the number of confirmed COVID-19 diagnoses per 100,000 inhabitants over the past 14 days, dividing geographical regions into green, orange and red zones depending on the number of cases. The red zone is considered to be the most dangerous. It is assigned to the region if cases exceed 150 per 100,000 people over two consecutive weeks or if the number of cases is between 50 and 150 but with the proportion of positive tests being above 4% of the total number of tests. As of today, all of Belarus is considered to be a gray zone due to the lack of reliable information according to the ECDPC.
We estimated the number of cases per 100,000 inhabitants over 14 days in November for each of the regions in Belarus and identified which zone would be assigned to them on the ECDPC map. The average value for Belarus is 1,113 new cases per 100,000 inhabitants over 14 days (the minimum value of the 95% credible interval is 734 cases, the maximum is 1,677). The minimum value of the lower limit of 95% credible interval is in the Homel region: 536 cases per 100,000 inhabitants, with the most probable value being 869 cases. The maximum value of the upper limit of 95% of the credible interval is 2,698 in the Hrodna region, with the most probable value in this region being 1,810 cases per 100,000. Thus, none of the regions of Belarus gets even close to 150 cases per 100,000 inhabitants over the two weeks. This demonstrates that the situation with the coronavirus in Belarus is at least no better and maybe even worse than in the most affected EU countries. For comparison, in Germany, the number of cases per 100,000 inhabitants over the two first weeks of November was 340, in Sweden 896, in Lithuania 1,053, and in Finland — 106.
In terms of regions, in absolute numbers, Minsk is the leader due to its larger population, while the Hrodna region has the most cases in percentage terms. Unfortunately, the lack of any publicly available information from the Ministry of Health makes it impossible for us to identify the reasons for this difference — whether it is due to wider availability of testing or a larger outbreak of the disease that may be associated with the circulation of a slightly different strain of the virus in the western part of the country. It is also worth noting that the leaked data that was published in the Telegram Belhalat_by channel confirms that the share (of the population of the corresponding region) of those who have been sick in the Hrodna region is the highest.
Cases of reinfection were estimated to be at the level of 1.3% on average (95% credible interval between 0.5% and 3.6%), while the number of those who think that they were sick twice but did not have a diagnosis is 14.4% (95% credible interval between 8.5% and 23.3%).
We are unable to assess the rate of mortality from coronavirus in our study based on the survey data. However, based on the responses of those who lost people in their social circle we were able to make a comparative analysis of mortality per month and conclude that more people died from coronavirus in October than in May. In May an average of 16% of adult citizens over 18 years old had someone from their social circle die of COVID-19 with the percentage reaching almost 25% in October.
Personal protective equipment (PPE)
How are Belarusians trying to protect themselves during the epidemic? Nationwide, people now wash their hands more often than before the epidemic (76% on average), use antiseptics (67% on average), wear masks or respirators in public places (54% on average). Besides, they try to limit contact with people (51% on average) and the use of public transport (40% on average). An average of about 23% of adult citizens wear a protective mask at all times. The most unpopular method of protection among adult citizens is always wearing a respirator. It is preferred by less than 5% of the country’s adults.
A week before the onset of the illness adult citizens did not actively observe safety measures. Those who had been sick were generally less likely to wear a mask or respirator in crowded places (an average of 46%), wash their hands (an average of 70%), limit contact with people (an average of 49%), and use public transport (average 37%) than all the others on average. To an extent, this demonstrates that the risk of getting sick is somewhat higher among the people who ignore these safety measures. This risk and its significance will be studied in more detail in our upcoming scientific publication.
At the same time, in the week before getting sick an average of 70% of adult citizens were using an antiseptic and an average of 30% were always wearing a mask. To some extent, this indicates either the ineffectiveness of such protective equipment or its misuse (which, in turn, creates a false sense of safety). For example, if one is always wearing a mask without changing it regularly, it will get wet and not only lose effectiveness but may increase the risk of getting infected.
It is worth noting that after being sick Belarusians generally begin to use personal protective equipment and wash their hands more actively, try to use public transport less, and have less contact with people.
The most common symptoms are fever, loss of smell (in approximately 70% of people), fatigue, muscle pain. However, most symptoms, except for fever, diarrhea, and dry cough, are more common in women. Interestingly, as age grows, the loss of smell becomes less common, whilst for older people it is more likely to experience fever and the loss of appetite.
According to our study among patients with a confirmed COVID-19 diagnosis, an average of less than 5% were asymptomatic. Overall there are probably many more people like that in the population but our data does not allow us to estimate this number.
Duration of the disease
Of those who have recovered, 43.5% had COVID-19 for a period of two weeks to a month. Nearly 30% recovered within two weeks, and just 5% were ill for less than a week. Generally, fewer than 2% of adult citizens were ill for more than two months. About 18% of all adult citizens who got sick have still not recovered as of 18 November.
Places of infection
The majority of adult citizens who have recovered, almost 45%, believe that they got infected at work. The second most common place the adult citizens believe they got infected in is home (family) — about 25%, while almost 20% are unsure where they got infected.
Only 30% of people who have recovered are satisfied with the quality of medical help provided, the rest had to face a long waiting time for examinations and test results (about 28%), had to pay to obtain a diagnosis out of their pocket (about 19%), were refused the PCR test (14%), computer tomography (13 %) or antibody test (11%).
Most of those who had been ill were treated at home (about 78%). About 15% of adult citizens were treated at the hospital, about 2% were in intensive care. We see that with growing age the illness becomes worse: generally, almost 30% of the adult citizens over 60 end up in the hospital and almost 5% in intensive care. Notably, some of the patients did not seek medical help at all.
Diagnoses were most often obtained using a PCR test — in 67% cases. Other methods include an antibody test (34%), lung CT scan (18%), and clinical symptoms (21%) were used for diagnosis. Note that these options are not mutually exclusive. In some cases, more than one method was used.
Fear of getting infected
If we compare the level of fear of coronavirus in the first and second waves, we see that on average the fear of being infected has remained at the same level: the higher the age, the smaller the share of those who have become less afraid and the larger the share of those who are just as afraid or have become more afraid. Women are afraid of getting infected more often than men.
Attitude to vaccination
The majority of the population (about 60% on average) are ready to get vaccinated if the vaccine is developed by a reliable pharmaceutical company, approved by the WHO, and has no serious side effects. On average just 4% of the population are ready to be vaccinated with a vaccine approved only by the Ministry of Health of Belarus.
Women more often than men believe that they will not get vaccinated under any circumstances and are in general more skeptical about coronavirus vaccinations than men.
Attitude towards the official position of the state on the situation with COVID-19
An absolute majority of the population does not trust the data reported by the Ministry of Health. Only a small part of adult citizens (up to 5%) consider the measures taken by the government to be sufficient. Most believe that they need to be seriously reconsidered:
- On average 80% of the adult citizens would like to see an increase in the availability of COVID-19 testing for everyone.
- On average 80% and 66% of the adult citizens want to see honest and detailed data on morbidity and mortality, respectively.
- An average of 66% believe that wearing masks in public should be mandatory. The survey was conducted before the mandatory mask requirement was introduced.
- An average of 62% would like to see more protection for the most vulnerable population groups.
- An average of 60% believe that educational institutions need to switch to distance learning.
- An average of 60% believe public awareness about issues surrounding COVID-19 needs to be raised.
Interestingly, the share of Minsk city residents who support the introduction of the measures and are of compliance with these measures is 5 to 10% higher than in the rest of the regions. Moreover, the older the adult citizens, the less important it is for them to know the real numbers on COVID-19 morbidity.
We have only included in this report the results that seemed most important. Shortly, however, we will start working on a scientific article where, in addition to the already published results, we will describe in detail the conclusions we reached based on our research and those that were not included in this publication.