(20b) Coronavirus statistics: how to present data about cases and mortality

An article and video to explain the statistics behind the coronavirus deaths but also to record illnesses.

 

Excess mortality

I'm in favour of using excess deaths (excess mortality) as a starting point to determine the severity. Of course, it doesn't need to be only death as the seriousness of something can also be defined as numbers of people who can't perform normal activities as is done for flu or heatwaves. But, as we're in the middle of a coronavirus pandemic that kills, I'll continue with excess mortality i.e. how many people die more compared with the average of the same months in previous years (see figure 1 for explanation). This excess death (and if data is available excess sick) informs quickly whether a day, week, month and year is normal or not.

 

Already this is done in the healthcare sector when GPs and/or hospital doctors notice an increase in patients and thus record the numbers of patients who are very ill or die to know whether a warning should be given. An example is a heatwave whereby we need to be sure people and certainly young and older people drink sufficient amounts of water and remain in the shadow. But, as the coronavirus crisis showed, in the worst hit regions many people die at home, either suddenly or hospitals are full and in such cases it is important to count all deaths to understand the urgency for actions and solidarity. Then countries can compare and explain difference whereby it seems eastern Europe acted more decisively than western Europe with significantly higher mortality in western Europe compared with eastern Europe for which a number of reasons are given.

 

Of course, it is also important to know whether people die from this coronavirus, another illness or simply of old age to know the severity of this particular virus. It seems certain deaths are unrelated to the coronavirus; then the question is whether they were tested for instance for the fly virus? And, as we're slowly approaching summer, during a heatwave more people and again mainly elderly die, it is important to know whether a higher number of people die because of the heatwave or from the virus so we know whether the virus continues during the summer or goes silent or even may disappear completely. Thus, even in the case we can't test all persons due to limited tests, it may be useful to collect blood and nose swaps at the time of death so afterwards the reason for people's death can be determined to complete the information and know better what happened. Further, concerning unconfirmed deaths, is it possible the virus may be undetectable after people survived the initial infection (with or without symptoms) but the virus caused sufficient damage so people die at a later moment when the virus is no longer detectable? Thus, have these people antibodies against the virus? Or damaged organs? And, we still need to know whether antibodies remain after months or may disappear.

 

Figure 1: What is excess mortality?

Statistics on affected groups

This shows that in addition to excess mortality also more detailed statistics are important to determine who are mainly affected such as age (younger versus older and what is the age when differences become apparent), gender (men versus women but even transsexuals may help to determine why morality may differ, i.e. is the difference a result of genes or more due to differences in levels of hormones), race (are differences due to genes or cultural differences such as closer contract or due to inequality whereby poorer people eat less healthy and thus are more obese and in general have more health problems that are known factors for a bad outcome), twins and closely related family members (to study possible genetic and/or environmental influences), ... .

Figure 2: To understand many articles, knowledge between graphs with linear or logarithmic scale is important

Belgium
Belgium seems to be very severely hit and is number one if we look to the numbers of people who died per million inhabitants per country (04/05/2020) although comparing countries is difficult as different methods are used. Why are morality so high as Belgium took quite early actions: early March first reported transmission, by March 12 actions by government such as closure of its borders, of shops, schools, bars and restaurants and non-essential companies while people were told to work from home if possible as many did and even going for walks was limited to two to limit contact between people as much as possible. Other countries where slower to act or didn't even go into lockdown while numbers remain lower. There are a number of possible reasons for this high number compared with other countries, future research will clarify.

 

One reasons is that Belgium counts all deaths, confirmed or suspected coronavirus patients and includes deaths from hospitals, care homes, home, something certain countries still refuse to do but in the end can't escape when the total mortality numbers will be compared with previous years. As explained higher, including everyone can reveal much.

 

One likely reason is that many people returned from (ski) holiday in severely hit regions such as northern Italy and Tirol at the start of the pandemic when it was not yet declared a pandemic so people still travelled freely and didn't even know the virus existed. Belgium has also an older population. The combination of people returning from holiday who may have visited (grand)parents (in care homes) to talk about their holiday should be taken into account. Therefore, graphs that compare countries should not all start at the same point zero when the first deaths were recorded but at the actual date that the first deaths were recorded so it becomes clear whether they had early infections and thus didn't know much about the virus or had later infections and thus may have been slow to act.

 

A severely hit region is the province Limburg and Liège that celebrated carnival at the end of February, beginning of March, thus around the above holiday period whereby people come in close contact, when the virus probably reached the region as people returned from they ski holiday and thus not long before the lockdown.

 

Further, Belgium is a densely populated small area. And thus, a few Belgians returning from holiday may be sufficient to spread the virus throughout large parts of the population, certainly as many people travel daily by public transport to the capital Brussels for work. Compare with large countries where a few cities may be severely affected such as NY in the USA because people travel within the city for work and pleasure while other regions may have fewer deaths so the overall numbers of death per million people can be relatively low. In contrast, when a small country was lucky, it may have no cases before a lockdown was imposed so no-one dies or becomes even ill from the virus.

 

Spain didn't include care home and unconfirmed coronavirus mortality in their statistics and I think that's not very wise. Indeed, what is not in the statistics didn't happen. Belgium took the decision to include all suspected deaths in its statistics, explaining its high mortality. Still, at a certain moment there was panic when the numbers of care home deaths rose and thus the overall mortality, demonstrating the urgency to take actions to increase the safety of care home residences. The statistics didn't look great but afterwards actions were taken so the situation imposed. Still, it's strange in many countries few actions were taken to protect care home residents.

 

Include all deaths, confirmed or not - than start to collect details

I think the best way is to include all deaths because otherwise, at the end of the month or year, it will become obvious that the numbers of people who died was much higher than what was originally recorded and this may result in another shock over the seriousness but also result in people talking about a cover up.

 

In addition, no records means no actions are needed as policymakers may not even know what is happening. When not all deaths are counted and thus mortality underreported, other countries may decide the numbers of deaths is not alarmingly high and thus may delay direct help for the concerned countries, countries may not understand the urgency to close borders and countries may not want to help financially afterwards. Collecting only numbers of deaths are not sufficient as we need to know what causes the death of people, the virus - directly or delayed, other causes?

 

Openness can save lives and increase solidarity. But, there will always be persons who can only think about money. They can't show empathy for the high costs in human lives but instead may think about savings such as fewer pensions that need to be paid. In that case, why do we have excellent scientists and healthcare systems when we don't want older people?