Sensitivity Analysis on the 'Under age 40 deaths' Substack
Findings after adjusting for right truncation of data
In this recent Substack, using Medicare “jab and death” data from Steve Kirsch, it was found that a significantly higher proportion of those under 40 were dying in the first year after the jab (compared to the number of them dying the following year; in Year 2).
But because data stopped being gathered on 1 Feb 2023, and some of the individuals did not receive dose #1 before 1 Feb 2022, the days-to-death would show up with right truncation (deaths which may have occurred after a year in those people would have been missed).
While all data in the data set are already deaths, had the data kept on being gathered, perhaps there’d be more dying in Year 2 than was seen. To adjust the counts in order to accomodate missing data, 50 early deaths were taken out of the Year 1 deaths and put into Year 2.
Here is the reasoning behind that number:
The 5-digit numbers along the bottom are Excel code for dates, and the value of 44592 — sitting just to the left of the orange line — translates into the day of 31 Jan 2022.
Each bar represents the next 50 days. While it may be possible that more than 50 deaths are missing, the fact that 50 were removed from Year 1 and then put into Year 2 represents a 100-death swing in the numbers.
Here are the notes showing how, even after a 100-death switch, findings are still statistically significant:
[click image to enlarge]
The code in the top window shows that the original value of 321 deaths in Year 1 has been reduced by 50 while that same number of deaths were added to Year 2.
Cell N35 shows that, even after the shift, the lower bound of the 99% confidence interval around the proportion of deaths by year still indicates that COVID jabs increase risk of death.
Early-only Samples
The most legitimate change would likely be simply to restrict yourself to those with 2 years of follow-up time. When I merely removed the 50 who could not possibly have had time to accumulate deaths in the second year, the sample of deaths reduced to 411, but the relationship stayed significant.
The 99% lower bound of the proportion dying in Year 1 (365 days or less from the jab) was .594 — still much higher than the “even-split” proportion of 0.50 which would be expected if COVID jabs did not increase the risk of death.
On a more restricted (n=135) sample where everyone had been jabbed before April 2021, the 90% lower bound of the proportion dying in Year 1 (365 days or less from the jab) was .503 — still marginally higher than the “even-split” proportion of 0.50 which would be expected if COVID jabs did not increase the risk of death.
By definition, a lower one-sided 95% confidence interval would show the same result: COVID jabs increasing risk of premature death, with the increase being statistically-significant at alpha = 0.05 (5%).
A most-restricted sample (most conservative)
Looking at just those jabbed by 18 Mar 2021, there were 99 total deaths recorded, with 57 of them (58%) being within a year of getting jabbed, and 42 of them (42%) being in Year 2 after the jab — but, due to having lower total amounts of data, the confidence interval could not be narrowed enough in order to statistically “prove” that COVID jabs cause those under age 40 to die.
Approximately 50 to 100 more total deaths would have needed to be in that sample to narrow the interval down enough to prove that COVID jabs were causing early death.
Technical Note: 89% of the recorded deaths involved people jabbed prior to 1 Feb 2022 and the distribution had positive skew.