Random. Came across this recently on the CDC website. Besides the obvious question of them now excluding 93 weeks of data during the actual pandemic and swapping it with numbers from 5 and 6 years ago why are they even bothering to change it this late into the game almost two years later and just last month?
Random. Came across this recently on the CDC website. Besides the obvious question of them now excluding 93 weeks of data during the actual pandemic and swapping it with numbers from 5 and 6 years ago why are they even bothering to change it this late into the game almost two years later and just last month?
I don’t care enough to speculate and am definitely no expert but removing 93 weeks of data seems like an odd way of improving the accuracy of their estimates. Ha
“On November 17, 2021, the algorithm used to estimate excess deaths was modified to include 6 years of prior data, rather than 4, which have been used to date. Initial estimates of excess deaths used 4 years of data prior to 2020 in the algorithm to evaluate potential trends in expected and excess deaths. As excess deaths continued to be tracked into 2021, an increasing amount of data had to be excluded from the algorithm modeling, as weekly counts of deaths during the pandemic (February 1, 2020 to present) are excluded from the algorithm to estimate the expected numbers of deaths. As of November 2021, this results in approximately 93 weeks of data being excluded from the estimation. With the exclusion of these 93 weeks, there was insufficient data to accurately estimate the trends in the expected numbers of deaths when using only 4 prior years of data. Thus, the number of prior years used to estimate the expected numbers of deaths was increased to 6, to ensure a sufficient number of years to estimate the trends over time. This change resulted in an increase in the weekly expected numbers of deaths by an average of 2% throughout the pandemic. The increase in the expected numbers of deaths resulted in a decrease in the estimates of total excess deaths, but these estimates are within the range of previously published values.
In previous releases, a range of excess death estimates were provided overall and for each jurisdiction, based on the total numbers of deaths above two thresholds (the average expected number and the upper bound threshold). With the improved estimates of the average expected numbers of deaths based on 6 prior years of data, a single estimate of total excess deaths is now shown for the US and each jurisdiction. Relying on the upper bound threshold to estimate total excess deaths will likely result in an underestimation of total excess deaths. The upper bound threshold is shown in the visualizations to illustrate when excess deaths are significantly higher than expected.”
Random. Came across this recently on the CDC website. Besides the obvious question of them now excluding 93 weeks of data during the actual pandemic and swapping it with numbers from 5 and 6 years ago why are they even bothering to change it this late into the game almost two years later and just last month?
I don’t care enough to speculate and am definitely no expert but removing 93 weeks of data seems like an odd way of improving the accuracy of their estimates. Ha
“On November 17, 2021, the algorithm used to estimate excess deaths was modified to include 6 years of prior data, rather than 4, which have been used to date. Initial estimates of excess deaths used 4 years of data prior to 2020 in the algorithm to evaluate potential trends in expected and excess deaths. As excess deaths continued to be tracked into 2021, an increasing amount of data had to be excluded from the algorithm modeling, as weekly counts of deaths during the pandemic (February 1, 2020 to present) are excluded from the algorithm to estimate the expected numbers of deaths. As of November 2021, this results in approximately 93 weeks of data being excluded from the estimation. With the exclusion of these 93 weeks, there was insufficient data to accurately estimate the trends in the expected numbers of deaths when using only 4 prior years of data. Thus, the number of prior years used to estimate the expected numbers of deaths was increased to 6, to ensure a sufficient number of years to estimate the trends over time. This change resulted in an increase in the weekly expected numbers of deaths by an average of 2% throughout the pandemic. The increase in the expected numbers of deaths resulted in a decrease in the estimates of total excess deaths, but these estimates are within the range of previously published values.
In previous releases, a range of excess death estimates were provided overall and for each jurisdiction, based on the total numbers of deaths above two thresholds (the average expected number and the upper bound threshold). With the improved estimates of the average expected numbers of deaths based on 6 prior years of data, a single estimate of total excess deaths is now shown for the US and each jurisdiction. Relying on the upper bound threshold to estimate total excess deaths will likely result in an underestimation of total excess deaths. The upper bound threshold is shown in the visualizations to illustrate when excess deaths are significantly higher than expected.”
https://www.cdc.gov/nchs/nvss/vsrr/covid19/excess_deaths.htm