Pediatric COVID Hospitalizations in the US

Data shown is as of 9/17. The HHS data file changed, and I need to re-write my import process. I will hopefully have it back up soon!

Frequently Asked Questions

What are the two different numbers for confirmed and suspected COVID?

Here are the definitions for the first two columns, directly from HHS for this dataset.

  • Pediatric Confirmed COVID – “Reported patients currently hospitalized in a pediatric inpatient bed, including NICU, newborn, and nursery, who are laboratory-confirmed-positive for COVID-19. This include those in observation beds.”
  • Confirmed & Suspected COVID – “Reported patients currently hospitalized in a pediatric inpatient bed, including NICU, newborn, and nursery, who are suspected or laboratory-confirmed-positive for COVID-19. This include those in observation beds.”

What does % Confirmed COVID mean? How are Louisiana and Nevada over 100%?

The third column is my own calculation, not part of the HHS dataset. This is simply the first column divided by the second. It shows what percentage of the total pediatric COVID hospitalizations in a state are patients with lab-confirmed COVID. Theoretically this should be almost 100% as hospitals can run tests very quickly now, but some hospitals seem more likely to consider patients with respiratory illnesses as suspected cases even after a negative test.

Louisiana and Nevada are routinely over 100%, because they seem to report to HHS incorrectly, where the confirmed COVID is larger than the confirmed and suspected number. I’m not sure if they are reversing the two fields, or if they are reporting only suspected in the second column (instead of the total of confirmed PLUS suspected).

What age is considered pediatric?

Pediatric includes children under the age of 18, in other words, 0-17 years old.

Is this data accurate? (Why did my hospital / local news / public health official give numbers very different than the ones shown above?)

Hospitals are required to report this data to HHS where it is compiled and provided in daily datasets. You can access the raw HHS data here (it’s easy to open in Excel – it’s just got one line per state). Sometimes numbers may get submitted wrong due to a typo or other error, but sometimes numbers are confused or exaggerated in the media. If you hear media reports that differ greatly from these numbers, here are some possibilities to consider:

  • Are they talking about CURRENT hospitalizations or NEW hospitalizations? The numbers above are for how many children are currently in the hospital. The CDC also provides data on new COVID admissions by age group, as a daily average over the past week (to the left of the graphs). It’s sometimes reported as a weekly total of new admissions, so you need to take the current 7-day average from that page and multiply by 7. Note that the total number in the top-left is all hospitalizations since August 1, 2020 (last year).
  • Are they talking about CONFIRMED or confirmed AND SUSPECTED? These are commonly confused, and the media loves to report the higher number that includes suspected cases without a positive test. That’s why I provided both numbers above for comparison. Nationally, about 1/3 of pediatric hospitalizations for COVID are only suspected (not confirmed).
  • Are they referring to very full children’s hospitals and pediatric ICUS, without specifying they are COVID patients? It is true that some children’s hospitals are very busy right now. With the surge in RSV, many children’s hospitals are strained right now, but that doesn’t mean they’re overflowing with COVID patients. They’re busy because kids get other respiratory diseases too. At Children’s Medical Center of Dallas, they recently reported they had 42 patients out of 44 staffed beds, but only 4 of those children had COVID.

I encourage anyone who is concerned that numbers might be wrong either in the media or in the HHS file to reach out to the appropriate contacts and ask for clarification. There have been multiple times where these numbers had to be corrected in the media after being questioned by ordinary people on Twitter (for example, in Mississippi and Texas).

Does this include children hospitalized WITH covid, but not DUE TO covid?

Yes, it does. The definitions of confirmed COVID hospitalization includes any children hospitalized with a positive COVID test, regardless of the reason for their admission. Previous studies have shown about 40-45% of pediatric hospitalizations are incidental to COVID, such as children hospitalized for trauma, surgery, obstetrics, psychiatric needs, and other conditions (California studies, CDC MMWR). These incidental hospitalizations still have to be isolated, so even if the child doesn’t need treatment for COVID, they are added strain on the hospital due to the COVID diagnosis.

For the younger ages (under 5), in the South and some other parts of the country, we are seeing a recent out-of-season surge in RSV, another respiratory virus that can be quite serious for young children. Many of these RSV patients also test positive for COVID, although the RSV is the likely cause of their severe symptoms. At Texas Children’s, they recently reported that over half (25 of 45) of their COVID patients also had RSV.

How does this compare to flu and RSV hospitalizations for kids?

The CDC estimates that there were 46,000 pediatric flu hospitalizations in the 2018-2019 flu season. The CDC uses estimates for flu because hospitals don’t test everyone for flu, and they aren’t required to report flu hospitalizations like they test for and report COVID hospitalizations. The CDC estimates that RSV causes an approximately 58,000 hospitalizations annually in children under 5 years old. According to HHS data, in a little over a year (8/1/2020-8/17/2021), COVID has hospitalized 49,000 children, which includes a lot of children hospitalized with COVID, and not due to COVID (as explained above).

Where can I see this data graphed over time?

Dr. James Salemi, an epidemiologist in Florida, has created visualizations for Currently Hospitalized adult and pediatric data using a similar HHS dataset to the one I’m using, but that is updated weekly and includes a time series. He also has a page for New Hospital Admissions by Age data, also based on an HHS dataset that is updated weekly.