The CDC's Flawed Data Tracker

The CDC Data Tracker Demographics page continues to grossly overstate pediatric deaths.

I was furious this afternoon when someone sent me this tweet from MSNBC reporter Mehdi Hasan claiming that there were “2,000 dead kids” from Covid.

Mehdi Hasan @mehdirhasan

2,000 dead kids, and 10s of 1000s hospitalized, was clearly not enough for a lot of folks in this country who would have preferred no protections whatsoever for our kids.

Ben Shapiro @benshapiro

You shut down schools, killed small businesses en masse, forcibly masked our kids, tried to force businesses to vaxx their employees, banned us from travel, kept us away from our dying relatives...all based on science you KNEW was shoddy. F*** you. https://t.co/GaHVN3xu2r

I track CDC’s official pediatric death numbers closely for my web site on a weekly basis, so I knew right away that number was not the accurate count. It comes from the CDC’s flawed Data Tracker Demographics page, which has a long history of overstating pediatric deaths (while oddly understating adult deaths) when compared against the official death certificate counts from CDC.

Kelley K @KelleyKga

This is a lie. The CDC says as of today there are 1,460 children with Covid listed anywhere on their death certificate, and for about a third of those, Covid is not listed as the underlying cause of death. cdc.gov/nchs/nvss/vsrr…

Mehdi Hasan @mehdirhasan

2,000 dead kids, and 10s of 1000s hospitalized, was clearly not enough for a lot of folks in this country who would have preferred no protections whatsoever for our kids. https://t.co/U3QsPsCxKZ

This Data Tracker problem is yet another Covid and kids issue that just keeps coming back to haunt me, so I found it a bit ironic that today is actually Groundhog Day.

Some background…

I discovered the CDC’s Data Tracker Demographics page in May 2021 when tracking down the source of an inflated pediatric death number that fellow Twitter user @EWoodhouse7 spotted in the media. At the time, the Data Tracker was overcounting pediatric deaths by about 200 (or 71%) when compared to the official source of CDC death data from the CDC’s National Center for Health Statistics (NCHS). The data from NCHS is is based on death certificates where Covid is listed anywhere on the death certificate.

Wood House 🔥 @EWoodhouse7

1/ There's a SIGNIFICANT discrepancy btw the number of pediatric COVID deaths the CDC is reporting on two different pages (h/t @KelleyKga for helping find it). A 204-death difference among ages 0-17 can't possibly be chalked up to "lag." 🧵

She and I both contacted the CDC about this discrepancy, and we have been fighting with them on and off about it ever since. After our emails in May 2021, they said they were looking into it, and they eventually made an adjustment to the Data Tracker numbers, but the Data Tracker pediatric deaths continued to grow at a much quicker rate than the official numbers from NCHS. In February 2022, I revisited the issue after seeing the inflated numbers being used again in the media and by Covid “experts” on Twitter, to scare parents about kids and Covid. At that time, the CDC’s numbers were overstating pediatric deaths by 61%, while understating adult deaths by 14% compared to NCHS data.

Kelley K @KelleyKga

2/ Despite being told in May 2021 that they were working to fix this, the issue still has not been resolved. Pediatric death numbers on the Data Tracker are 61% higher than NCHS data, while over all ages the Data Tracker is 14% lower than NCHS.

After contacting the CDC directly, and getting journalists involved to help ask some questions, the CDC made a big correction, blaming the problem on some bad computer code. The BMJ even published an article about the flawed data and the revision. However, I don’t think the CDC actually fixed anything, because the problem has continued where the Data Tracker grows much more quickly than the NCHS data — just for kids.

What are the real numbers?

There are two official ways that the CDC tracks Covid deaths: from surveillance data summited by state health departments and from actual death certificates. The Data Tracker is based on the surveillance data, while the NCHS data uses death certificates.

Within the NCHS data, there are actually two ways to report Covid deaths. The NCHS “Weekly Updates” page shows “All Deaths Involving COVID-19” (emphasis mine). These are deaths where Covid is listed as the Underlying Cause of Death OR a Multiple Cause of Death (typically for Covid, this means Covid is listed as a “contributing” factor). Here’s how the Data Tracker compares to the NCHS deaths as of last week:

Kelley K @KelleyKga

The Data Tracker hasn't updated yet for this week, but here's how much the numbers were off last week from CDC's official source. (NCHS numbers listed include Covid as a contributing cause.) Adults UNDERstated by 14%, pediatric OVERstated by 41%. How is this so broken?!?

The Data Tracker is overstating pediatric deaths by almost 600, which is 41% of the NCHS total. Meanwhile, it continues to underreport adult deaths by 14%. Why?!?

The CDC also has a system called CDC WONDER, where you can perform a search query for the number of deaths where Covid is listed only as the Underlying Cause of Death. This means the numbers from the CDC WONDER database are even lower, since they exclude deaths where Covid was only a contributing factor. WONDER updates monthly, and is due for an update any day now, so it’s a bit behind, but you can see it currently shows a little under 1000 pediatric deaths where Covid is listed as the Underlying Cause. (I will update this screenshot when WONDER data updates next.)

Source: CDC WONDER Online Database

Three Years Worth of Data

As you can see above, while people typically refer to deaths from cumulative Covid deaths, it’s also important to remember we’re going on three full years of dealing with this virus, and several different waves. The annual numbers are much smaller, and the highest numbers seem to be behind us (over half of the 2022 deaths occurred in January - March). For comparison, the CDC estimated there were 1,090 pediatric deaths from the H1N1 flu epidemic, from April - November 2009. (My daughter was an infant at the time, and I don’t remember being worried about H1N1 at all. She went to Mom’s Morning Out at our church, she crawled around the play area at the mall, we flew across the country twice for family weddings...)

---

Pediatric deaths from Covid truly have been rare, and it’s just absolutely disgusting how some supposed experts and the media frequently exaggerate and twist the data at every opportunity to terrorize parents. It’s cruel. I hate that US public health has been so determined to use fear about children and Covid as a tactic to push vaccines and other mitigations.

The CDC knows their Data Tracker Demographics data is not accurate, not consistent with their official data, and grossly overstates pediatric deaths. But they seem unwilling or unable to fix the data. And because it’s easy to access on the CDC’s web site, it’s routinely cited by experts on Twitter and in the media. Despite repeated notifications about this data being wrong, the CDC just continues to update the page with more misleading data. At this point, one has to wonder if it’s intentional.

Phil Kerpen @kerpen

@KelleyKga Tracker's sole reason to exist, as far as I can tell, is to propagate this lie.

I’ve contacted the CDC through their web form about updating the Data Tracker Demographics page, and I tagged some journalists on Twitter who I thought may be more successful at reaching out to the CDC on my behalf to get this data corrected (again). But I really wish they’d just remove this page completely, since they seem unable to ensure it’s accuracy. Unfortunately I don’t see that happening.