How Steve Kirsch is getting it wrong about the vaccines

Can I win a cash prize for this?

On September 1st, 2022 Steve Kirsch published a post with the slug “this-one-graph-tells-you-everything” and with the title “Vaccines are taking an average of 5 months to kill people”. 

I submitted a brief ‘steel-man’ submission in his contact form that included a blue/yellow graph from my colleague Herve Seligmann. The graph shows positive (yellow) correlations between vaccination rates and mortality 0-5 weeks post-injection, negative (blue) correlations between weeks 6-20, and then positive correlations again starting around 24 weeks. However, I suggested there is an alternative explanation for the second yellow hump in Herve’s graph as well as for Steve’s insurance graph.

Here, I flesh out why I believe that Steve’s insurance graph does not reflect delayed deaths from the jab, but rather it reflects COVID associated deaths due to a delta wave in the (mostly southern) US states that began in early August and peaked in mid-September, 2021.

Steve’s initial basis for his conclusion was a graph from the US Social Security death master file showing a peak in mortality around September 9th, 2021: 

He paired this with a graph of vaccine doses administered across time from Our World in Data which shows the most doses given on a single day was 3.5M on April 11th, 2021.

The difference between the two peaks (5 months) is what Steve suggests is evidence that any fatal consequences of the vaccines take an average of 5 months to kick in.

However…

It is important to note there was a delta COVID wave ripping through the US (mostly southern states) during the time frame of the above insurance graph. The peak of the COVID wave occurred around Sep 21, 2021 as shown below.  

So, one would have to first rule out COVID-associated deaths if we are going to believe that the insurance graph reflects any delayed fatal consequences of the vaccines. Steve states that “In addition, the insurance companies confirmed it wasn’t COVID that was causing the excess deaths”, but it would be nice if we could verify or confirm this for ourselves.

One way to explore this would be to extend the insurance graph to begin earlier in the year, before widespread vaccination, to see if the same plot shows the previous COVID wave which peaked in the first weeks of January, 2021. If it did, that might suggest that COVID was in fact contributing to the deaths reflected in the insurance graph.

Unfortunately the insurance graph starts from April 1, 2021, and the underlying data used to generate it is not publicly available so it will be impossible to extend the graph directly.

So, I tried to recreate that plot myself using publicly available data. I pulled this publicly available spreadsheet from the CDC which tracks weekly deaths by age from 2015 to the present (thanks to Josh Guetzkow who initially brought it to my attention back in December, 2021). 

https://data.cdc.gov/NCHS/Weekly-Counts-of-Deaths-by-Jurisdiction-and-Age/y5bj-9g5w 

I created a similar version of the insurance graph with the exception that it compares percent changes (2021 vs. 2020) in weekly deaths for ages 25-64 (the spreadsheet stratifies deaths according to the following age groups: <25, 25-44, and 45-64, and >64).

We see that the second hump corresponds to the insurance graph with a peak around weeks 33-34 (first half of September), while the first hump corresponds to the previous COVID wave that peaked around the 2nd week of January.

The below graph plots the same for ages to 25-44 and shows a similar pattern.   

All the above suggest that the insurance graph is showing COVID-associated deaths, not necessarily delayed deaths from the jab. 

If we extended that insurance graph to earlier in the year, would we see a similar peak? It would be great if Steve could get the insurance graph for the whole year.

Since Steve put up his initial post, he has edited it to include additional analysis and seem to support his claim. I will aim to address those analyses and see if I can find alternative explanations for those as well in the near future. Stay tuned!