Pandemic raging, yet fewer Hungarians die than before coronavirus. But why?
As far as data in the current coronavirus pandemic go, excess mortality is considered both in Hungary and internationally as the most reliable 'non-coronavirus' statistics. It shows the increase (or decrease) in the number of deaths during the weeks of the pandemic over the same period of the previous year(s). Many hold this particular statistics in high regard because it simply circumvents the various definitions of who can be considered as a 'coronavirus victim', how many people with coronavirus infection are detected, or if everyone in hospitals are tested for COVID-19 or not, and so on. Instead, it simply compares the number of deaths in a given period of the pandemic with the number of deaths in a corresponding period when SARS-CoV-2 was still nowhere.
Hungary's excess mortality developments are clear, as attested by the chart below. Hungary had a mortality deficit in early 2020, before the coronavirus outbreak, i.e. fewer people died than the 2015-2019 average on the same weeks. They say the reason was the less deadly flu season. The spring wave of the pandemic did not lead to many deaths due to the rapidly implemented lockdown measures therefore the demographic changes up to the autumn were not out of the ordinary. The second wave, however, brought about extremely negative changes and a lot more people died than in the same weeks/months of the previous years.
In fact, almost all of the heavy losses were suffered over a 15-week period. But then the number of weekly deaths started to fall quickly and got below the multi-year average fast.
More in the article:
- How many more people died in 2020 than what official data show?
- How come this year's excess mortality is so low?
- Why does the statistical office 'find' thousands of new deaths eventually?
Comparing excess mortality to the official number of people that died of coronavirus-related diseases, we can draw several thought-provoking conclusions. As you can see on the chart below, the two indices do not move on the same trajectory. In fact, there are seemingly unexplainable differences between them.
The two most prominent phenomena:
- At the peak of the second wave, Hungary logged 1,100 new COVID-19 deaths per week, while excess mortality was well north of this over 1,500. During the upward trend of this wave, 4,700 more people died than what official coronavirus stats explained.
- And then came a turnaround at the end of 2020, a most spectacular one at that. The unexplainably high excess mortality switched into a mortality deficit. This shows that over the past ten weeks (we have no fresher data at this point) 7,700 fewer people died in Hungary than what would stem from regular demographic changes and the impact of the coronavirus pandemic. To put it differently, regardless of the more than 500 lives claimed by the pandemic every week between the second and third waves, up to 250 to 300 fewer people died per week compared to normal years.
The odd phenomenon indicates that excess mortality statistics are not as almighty as we might think. Not only is it rather 'slow', it should also be applied to gauge the gravity of the pandemic only with great circumspect. Secondly, the low mortality figures recorded this year beg the question: is coronavirus less deadly than what previous experience, researches and studies suggested? We guess it's not a huge spoiler to reveal that, unfortunately, this is not the case. So let's take a look at all the static regarding excess mortality!
Statistics
The Central Statistical Office (KSH) publishes the number of deaths every week. The release has recently been brought forward, but the usual price had to be paid. In exchange for speed, the accuracy slackened. Up to the autumn of 2020 this meant that every time the KSH published its latest figures, it also revised its data for the preceding two weeks upwardly. And the revisions went like clockwork. On average, 2,500 deaths climbed 100 higher upon the revision.
This changed entirely in the second wave. The revisions were greater and were done retroactively not for two but for several weeks. The number of weekly deaths often rose 300 to 400 compared to the first estimate, which marked an adjustment of around 10%.
We have inquired at the KSH about this change. It said the more drastic revisions likely had to do with the higher number of deaths leading to delays in the registration of deaths. In some weeks last year, excess mortality was 60% which must have put an extreme load on the issuance of death certificates, as well. That is why it was unexpected that the President of the KSH announced out of the blue on Monday that they have closed the 2020 mortality statistics and these are the final figures. (It's unclear what Gabriella Vukovich meant by 'final', as the data series clearly show (numbers with blue are not final) that revisions will be made.)
If we add the weekly revisions by the stats office since the start of the second wave we get nearly 4,500 people. So, what does this mean from our point of view in respect of the odd behaviour of the excess mortality statistics? It primarily implies that the excess mortality (actually mortality deficit) numbers of the last few weeks will be revised upwardly. The extent of the revision is uncertain, as the rate of the corrections have been moderating over the past few weeks, so the columns for the 4th to 7th weeks on the chart above may not rise to over 300. However, it is possible that the mortality figures for early 2021 will not be lower than the 2015-2019 average.
The influenza
There must be some explanation other than coronavirus to the 500-strong weekly excess mortality of the second wave and the 'lack' of 700 deaths per week on record since. As for the former, we might never have a satisfying answer. Last year's high mortality data (that continue to rise due to the revisions) are also surprising because under the local methodology, everyone that was once diagnosed with COVID-19 and died later (no matter how long after their diagnosis) are considered as a victim of coronavirus. Still, the answer might be that a lot of the deceased who died of coronavirus-related diseases during the second wave were not registered as COVID-19 victims.
This assumption is underpinned by reports from the period about hospitals being full of patients with severe coronavirus symptoms but no positive tests due to the shortage of testing capacities. If that was the case, the actual number of people that died of coronavirus-related diseases could be four to five times higher than what the official numbers show.
Others looking into the data, as well
While we were working on this article, the Hungarian Demographic Research Institute has also published a report. The main finding is that excess mortality reached 13,700 between the 12th and the 52nd week in 2020. In other words, this many more people died during the coronavirus pandemic last year than how many would have died without the outbreak. Thsi figure is more than 4,000 higher than the number of COVID-19 deaths and is almost entirely in harmony with our estimates. The author, Csaba Tóth G., thinks the explanation of the difference lies in the narrowing capacity of the health care system and the extreme load it had to cope with.
What we should focus more now is the period that started when the second wave was apparently petering off. As we have already noted above, more than 500 people per week died between the second and third waves, yet there could have been 250 to 300 fewer deaths per week than in a normal period. In other words, there was a factor (the pandemic) which leads to 500 more deaths a week, yet the total number of deaths shows a decline over the previous years. This controversy may be mitigated by the upcoming revisions, but it will certainly not be eliminated. And that's because the key factor here is not this, but the flu.
The excess mortality index assumes (at least that is our assumption when we interpret it) that the difference between the average mortality of the base period and the number of deaths in the period under review stems entirely from the coronavirus pandemic. This means that if it was not for pandemic there would be this fewer deaths in Hungary. However, if we take a look at the first chart we find that the number of weekly deaths is not steady: it is below 2,500 in the summer and reaches 3,000 in winter.
Winter months are generally flu season and the flu also claims lives, although a lot fewer on average than the current pandemic that the cabinet was/is trying to keep in check by various lockdown measures. Consequently, when we calculate excess mortality, our base for the comparison becomes higher and higher death tolls in non-pandemic periods as we approach the end of the year. However, there is no sign of a flu epidemic at the turn of 2020 and 2021. You'd need a microscope to find people who have the flu.
As we have pointed out in one of our previous analyses we should ask ourselves how on earth should we calculate/interpret excess mortality then? Well, it depends what we want to demonstrate.
If we want to find out how many more people die now than if there was no coronavirus, the above calculation appears to be right. Without COVID-19 there would be no lockdown measures in place, but the influenza would spread and it would kill people. Consequently, the restrictions make sure not only that fewer people die of coronavirus, but also that fewer people die of influenza.
If we want to find out (in a partial approach) how many people could have died in relation to coronavirus, we need to compare to a period when there's no influenza epidemic, either. This would result in higher excess mortality than what we have now. In order to illustrate this we show you how excess mortality would look like we compared every weekly reading to a steady weekly death of 2,500 people.
As you can see, most of the massive gap between excess mortality and COVID-19 deaths disappears, and the data revisions of the following weeks are likely to narrow this gap even further.
Another mortality weekly report is scheduled for publishing this week and we'll have all data for February that can prove us correct. But then we'll also have mortality data for the third wave of the pandemic that will present another conundrum as to how we should interpret the excess mortality indicator.
