Summit
Critical Crazy
- 2,702
- 1,323
- 113
The 0.2 per 100K is deaths (due TO COVID) vs total population (1 in 2M) the numerator (25) is the exact as the 5 per 100K stat. The denominator for 5 per 100K is dead d/t (not with) COVID over diagnosed. The 0.2 per 100K is dead over total population. That is the number I was calling out as meaningless at best because most people haven't had COVID-19 and their study time was Mar 2020-Feb 2021.
ETA That being said, COVID prevalence in CYP mortality appears undersized vs total seroprevelenence by a factor of 2-3 to 1 which of course tracks our knowledge of COVID being less severe in CYP but also here that it is relatively less severe than other causes and you actually do need to wonder about "of vs with," which is why the authors examined this exact thing (there's an algorithm in the study that explains how they did this) and they came up with at least 41% of those who died with COVID (61), died of COVID (25), with the rate being higher the older the patient was.
That metric is how you would examine an endemic disease for comparison across time to determine risk. That works pretty well for disease (infectious or not) with minimally/slowly changing incidence, morbidity and mortality.
It is nonsense for a pandemic with a dynamic novel virus in a largely naïve population when the virus is changing its virulence and R0 over timeframes of months and thus frequently goes exponential with curves that look like any Reed Frost until mitigating factors kick in.
Knowing when to use what method of analysis is basic epidemiology which gives me pause in examining this preprint, although I shouldn't impart intent on the investigators as it may merely be a statistic of interest, but really relevance of statistical points is something of great importance.
With the high R0 of Delta the vast majority, probably 70% or much higher, of unvaccinated folks will get COVID if they have not already had it. Given the discussion about vaccine these days revolves almost entirely around the individual's personal decision of whether or not they should get vaccinated, the question is most obviously NOT what is my risk based on a retrospective snapshot of the whole population, but what is my personal risk going forward: risk from vaccine and risk from disease.
Let me put it a different way: if there was incredibly rare brain cancer with 100% mortality for those who got it but it only killed 330 people a year in the US, the rate is 1 in 1 million because it is rare in incidence. Both numbers are interesting. But COVID-19 is something that almost everyone who isn't vaccinate or already recovered is going to get. So for those people, looking at the mortality rate across the whole population in a given timeframe is not useful to inform risk decisions for an individual or subpopulation looking forward.
Nobody is giving vaccine AE rates based on total population
That is NOT what I said at all in my post. I made a risk-benefit argument that an individual could use to make a voluntary decision for themselves or their child who is low risk because the vaccine is low risk because the available data shows disease to be both inevitable and higher risk than vaccine. You counter that you don't think there is enough data to back my position. Thus I ask you what would be sufficient?
Last edited:
