3 ms·
Sciensano is the science authority for Belgium. So I don't think there would be a lot of bias. The same agency suggested overcounting instead of potentially un
by NicoJuicy 5y ago
Sciensano is the science authority for Belgium. So I don't think there would be a lot of bias.
The same agency suggested overcounting instead of potentially undercounting, this way, not every death should be tested.
- tinus_hn 5y agoBias is everywhere and not necessarily malicious. But if you randomly select a percentage of all positive tests for checking for which variant it is, you’ll get a different result than when you randomly check a percentage of all people who end up in the hospital.
- NicoJuicy 5y agoThe numbers, according to the article/pdf, are people checked in the hospital in severe condition. Not the other way around
- tinus_hn 5y agoThe paper claims this is described in the ‘clinical hospital surveillance’ and although the report about that actually does not mention sequencing or Omicron there might be a clue as to why they don’t see data on omicron patients in ICU: https://covid-19.sciensano.be/sites/default/files/Covid19/COVID-19_Hospital_epidemiology_Part_1.pdf https://covid-19.sciensano.be/sites/default/files/Covid19/CO... Page 7: > Clinical outcomes, such as ICU admission or death, are only registered at hospital discharge […] > Furthermore, as ICU patients remain hospitalized for 2-3 weeks, their profile may only be available approximately 4 weeks after their initial hospital admission. To avoid misinterpretation of trends over time we do not report ICU data for the most recent 3 weeks.