Of course the NYTimes Used Bad Data
David Zweig sheds light on ANOTHER use of terrible CDC data
First, go subscribe to David’s substack. He was one of the very few journalists who brought integrity to his job these last 3 years.
Here’s what he found:
An article in Saturday’s New York Times includes the following passage:
This is false. Nearly 2,300 American children and adolescents have not died from Covid.
There are two problems here. First, the CDC knows this number is wrong, but it shares this number publicly anyway. Second, many journalists, including the three New York Times reporters on this piece, continue to report these incorrect numbers.
This is the umpteenth example of our public health agencies providing misleading or outright incorrect information and journalists reporting it without making an attempt to verify its accuracy.
Let me briefly explain the problem with the statistic given in the NYT article:
The CDC maintains two different publicly available tallies of Covid deaths. One is accessible through the Covid Data Tracker demographics trends page. The mortality numbers on this dashboard “include probable COVID-19 cases and deaths,” and are derived from data “reported by state and territorial jurisdictions to the Centers for Disease Control and Prevention.”
The way the NYT reporters got “nearly 2,300” was by adding the numbers from each of the pediatric age brackets, which totals to 2,281. The problem is these numbers are not correct, and the CDC knows they are not correct.
The CDC uses the Data Tracker to observe trends in real time. Its numbers are preliminary and rough. And they have consistently overcounted pediatric deaths.
The second pediatric tally by the CDC—which is significantly lower than the one presented in the Data Tracker—is through the National Center for Health Statistics, and it is the official death toll, compiled via death certificates. How do we know NCHS (not the Data Tracker) provides the official tally? Because the CDC said so:
In a statement to the journal BMJ, in a March 2022 article about an error with the Data Tracker, a CDC spokesperson said the Data Tracker numbers “are subject to change,” whereas the National Center for Health Statistics “is the most complete source of death data.”
In an email to journalist Stephanie Murray at the same time the CDC similarly said that the NCHS data are “the official death statistics” for Covid.
What is the actual total for pediatric Covid deaths? The latest figure from NCHS is 1,668.
Though there is a lag of a few weeks for death certificates, the lag does not account for the vast difference in pediatric numbers between the Data Tracker and NCHS. The pediatric numbers from the Data Tracker have been consistently wrong for years. Discrepancies between the two databases, and specifically the inaccuracy of the Data Tracker’s pediatric data, were pointed out as early as May of 2021—not by scientists or journalists, but by two regular people, Jessica Hockett and Kelley Krohnert. Krohnert has written extensively on this issue and CDC statistical errors in general.
It is also worth noting that although the NCHS number is the official number, even that number is for “all deaths involving Covid,” which includes those where Covid was not the underlying cause of death and where Covid was not laboratory confirmed.
Although the difference between the Data Tracker versus NCHS numbers may make some people’s eyes glaze over, it’s important to recognize that 2,300 is nearly 40% higher than the official total. These numbers matter. They are used to justify policy positions—both for the future, and to retroactively justify certain policies (such as school closures) that happened in the past—and they are the historical record.
If the CDC wants to use rough data for its own purposes to track trends then it should do so privately or have a clear disclaimer to the public that these are not official numbers. And journalists who cover Covid should be aware of the difference between the two databases, and they should not cite the pediatric mortality numbers from the Data Tracker.
Special thanks to Kelley Krohnert for letting me bounce this article off of her before publication.
I don't care how many children are said to have died of a cold. The truth is zero. They may have died of lack of nutrition, or medical mitigations, or not enough sunshine and fresh air and water and healthy salt, but no kids die from having a cold. Ever. Zero. You can not exaggerate zero.
I have absolutely no dispute with your article, your analysis, or the work of Jessica Hockett, who aside from being someone I am acquainted with, is pretty far from "an ordinary person." She is a data analysis maven, and (in full disclosure) a PhD as well. All that preface aside, even if the number was 2,300 pediatric deaths from covid, over the course of the pandemic, that number *still* does not justify closing schools. Something like 7K children between 1 and 14 die every year from the Top 10 causes of death in those age groups. Now, to be clear, and to prevent anyone jumping to a conclusion and pulling a hamstring, I am not saying deaths, of any cause, are "OKAY" or any such thing. However, again, the number cited by the NYT should not justify school closures, even if it were the correct number. At least, that is my current position. Happy to be convinced otherwise.