Pleidooi voor een ‘normsterfte’

CBS tries to predict as accurately as possible. In relatively quiet times, without disasters, wars or other attacks, that prediction could be used as a standard for approximately how many people would 'be allowed' to die in the coming years. If the actual mortality was above or below it, we spoke of excess and undermortality. We called that prediction the 'baseline', on the basis of which we could see whether there was too much or too little mortality. That is no longer possible.

Earlier I drew the parallel with body weight: we can predict what the average BMI will be in the coming years. However, that does not mean that it will become the new standard. Let alone that it is the desired situation. Even if more than half of the population is overweight, this does not automatically become the norm for a healthy weight. A different standard has been set for this¹https://www.voedingscentrum.nl/bmi which is separate from the annually observed values. If that BMI standard were not there, it would be impossible to claim that more than half of the population is overweight or that obesity is becoming an increasing problem.

How different it has been with mortality in recent years. Actually, we should be talking about mortality rates by age group and by gender and not about the mortality rates counted, because those figures are also influenced by factors such as changing population size (including migration) and age distribution (ageing). The fact that we see rising trends in the graphs below is mainly due to this. But how do we see that future? I have some CBS forecasts compared with each other.

In the forecast from 2008 we see a straight line (light grey). In the long term, that cannot be right because the country is too small for that. It should be a curved line, but the curvature is so slight in a period of less than two decades that it will not make a decisive difference to what we want to argue.

Forecasts compared

The CBS forecast from 2008 resulted in structural undermortality: the blue area does not even touch the predicted values every 6 years thereafter. The actual mortality level was considerably lower.

In 2014 zien we hoe de prognose inmiddels is bijgesteld. De groene lijn. Wat er op dat moment bekend was van 2014 werd al meegenomen (dat kunnen we in 2020 ook zien). Trek je die groene lijn naar links door, dan zit je perfect. Maar het zat niet mee: de daaropvolgende jaren was de sterfte weer te hoog en klopte weer met de oude trendlijn van 2008, terwijl 2015-2016 moderate flu years goods. So why not somewhere in between?

And indeed. The blue forecast was published in 2017. That seemed to go quite well, but then came 2020: Covid. A (very) strong flu. This was included 👇 in the 2020 'forecast' itself

Forecast-2020 was published in December 2020. But 2020 turned out to be even higher. It was expected that this would linger a bit in 2021, but after that you should see some undermortality again. The red line therefore falls below the forecast of 2017, towards that of 2014 (which had previously turned out to be too low). This resulted in an insane excess mortality. RIVM had to give it a twist. This was no longer a statistical problem, this was a health problem.

For 2024 and beyond, which is no longer in the graph, CBS does not expect a quick return to 'normal', whatever that may be. (I have put the total for 2024 in graph at 169,800, which is not definitive).

Mogen we ons afvragen waarom ze in 2020 wél dachten dat het weer snel zou afvlakken, en nu niet? Waarom dan niet…? Willen ze het RIVM niet tegenspreken of weten ze wat er aan de hand is?

In other words, according to CBS (in the latest forecast dated 2023), for 2023 and 2024 combined:

• There are 21,938 more deaths than what CBS had calculated in 2014.
• 16,993 more deaths than what CBS had calculated in 2017.
• There are 19,622 more deaths than what CBS had calculated in 2020.

That's 23 to 30 deaths per day MORE than according to previous calculations. What happened between 2020 and 2023 to justify an increase of almost 20,000 deaths? We don't know. Researchers from the Princess Máxima Center came up with some ideas, but were sharply rebuffed.

So the mortality expectation has suddenly risen significantly. What does CBS say about the increased mortality expectation via vzinfo.nl?

Annual mortality continues to rise, reaching around 210,000 by 2055
As a result of the aging population of the post-war baby boom, the number of deaths will continue to increase in the coming decades. While there were still about 170,000 deaths in 2022, it is estimated that in 2055 there will be about 210,000. From 2055 onwards, absolute mortality is expected to start declining again.

https://www.vzinfo.nl/sterfte/toekomst

Nothing at all about the skyrocketing expectation that has nothing to do with the baby boom. Are they perhaps saying something about Covid on that page? Yes:

In the corona years 2020 and 2021, there were approximately the same number of deaths among men and women.

https://www.vzinfo.nl/sterfte/toekomst

Not a word about the increase in mortality that we are in the middle of. The term "excess mortality" is hardly applicable anymore because both CBS and RIVM apparently take the unheard-of increase into account in their expectations. Statements? I paraphrase: "Probably Long-Covid. After all, it can't be anything else."

Wat er moet komen is een gezondheidsnorm, een ‘normsterfte’. Voorspellingen zijn zinloos. Plotselinge afwijkingen van de normsterfte, incidenteel of meerjarige trendbreuken, moeten verklaard zijn én bewezen onherstelbaar voordat de normsterfte hierop wordt aangepast. Hiervoor zijn criteria nodig want geleidelijke afwijking is onontkoombaar. Het zou zelfs iets zijn voor een onafhankelijke commissie die het ministerie van VWS adviseert.

Herman kan weer aan het werk! Sterker nog: hij IS al aan het werk. 😊

EDIT: De eerste respons op het pleidooi can be found here

Sources:

Go to https://opendata.cbs.nl/statline/#/CBS/nl/ and search "Prognosis of deaths" (with quotation marks)
2023: https://opendata.cbs.nl/#/CBS/nl/dataset/85753NED/table?dl=B15DC
2020: https://opendata.cbs.nl/#/CBS/nl/dataset/84882NED/table?dl=B15DD
2017: https://opendata.cbs.nl/#/CBS/nl/dataset/83794NED/table?dl=B1604
2014: https://opendata.cbs.nl/#/CBS/nl/dataset/82815NED/table?dl=B15E1

For previous forecasts:

2008: https://opendata.cbs.nl/#/CBS/nl/dataset/71867ned/table?dl=B1953 up to and including https://opendata.cbs.nl/#/CBS/nl/dataset/71867ned/table?dl=B1954

References

• 1
  https://www.voedingscentrum.nl/bmi