Thread by @Goldammerfeder

No. It’s NOT worrying. It’s scaremongering on your part.

The numbers displayed on the danish 🇩🇰 dashboard don’t reflect sufficiently the real situation in hospitals anymore.

That’s due to the way the data is collected.

Explanation below.

https://twitter.com/gab_h_r/status/1486326307034218501

In the official Danish data published on the dashboard by @SSI_dk, EVERYONE who tests positive within the 14 days before or on admission to hospital is included in the statistics.

Before #Omicron, SSI provided good real-time data in this way, very close to reality.

However, since Omicron, this is now becoming a problem. That's because infection rates are increasing while at the same time disease severity is decreasing (vaccines + Omicron)

If the infection rates then go as high as in Denmark, data collected in this way become very unclean.

Because according to the SSI definition, even someone who was PCR+ 12 days ago, is out of isolation again, then has a traffic accident and tests PCR- in the hospital is included in the statistics.

This issue is described in detail in this article:
nyheder.tv2.dk/samfund/2022-0…

So you have to distinguish 3 groups of hospitalized people in Denmark:

Group 1: tested PCR+ within 14 days before, but PCR- on admission.
Group 2: PCR+ on admission, but hospitalized due to other causes (with COVID)
Group 3: hospitalized because of COVID-19

Hospitals self-report significantly fewer patients they treat with/because of COVID (Group 2+3).

January 17/18 data:
Hospitals report ~40% fewer patients hospitalized with/due to COVID than SSI published as official numbers.

Dark red: hospital numbers
Light red: SSI numbers

In addition, hospitals reported on Jan 17/18 that of those hospitalized with PCR+, over one-third were hospitalized for another cause. COVID was thus only a secondary diagnosis.

dark red: total
red: because of COVID
light red: with COVID

Thus, of the patients shown on the SSI dashboard, on January 17-18:

Group 1 (tested negative in hospital): ~40%.
Group 2 (hospitalized with COVID): ~20%.
Group 3 (hospitalized because of COVID): ~40%.

The problem is well known, including SSI.
It is also clear that as infection rates continue to rise, the proportion of groups 1+2 will continue to increase, and SSI numbers are increasingly different from hospital reality.

SSI is trying to address this issue by now reporting how many of those PCR+ within 14 days of admission are in psychiatric units - because yes, these are also included in the statistics.

As of today: 24%
experience.arcgis.com/experience/aa4…

Of course, this does not provide a complete picture of the exact proportions of groups 1+2 in the total.
But the change in hospitalized in psychiatry are a good proxy of how the proportion of groups 1+2 is changing.

At launch on the dashboard: 18%.

https://twitter.com/mikaelmilhoj/status/1480974196163420163?s=21

Same problem with death rates, by the way - except there is a PCR+ window of 30 days.

@SSI_dk and @SSTSundhed are working on providing more accurate data that incorporates medical protocols/diagnoses. That's just not real-time data anymore and not provided via the dashboard.