Read on Twitter
This looks like a 4k word story about a database, but it's really about how the government *can* build capacity to battle the pandemic.
We didn't know which hospitals were in trouble. Now, we do. How'd that happen? https://www.theatlantic.com/health/archive/2021/01/hhs-hospitalization-pandemic-data/617725/
We didn't know which hospitals were in trouble. Now, we do. How'd that happen? https://www.theatlantic.com/health/archive/2021/01/hhs-hospitalization-pandemic-data/617725/
This story has been on my mind for many months as @HHSGov improved and published ever-more hospital data, especially given how much criticism rained down on them for moving hospital-data ingestion away from CDC.
Some of this is reasonable. Trump's HHS had some political hacks. The CDC has many great scientists—and people really want them to be the heroes in this pandemic.
But we realized after a couple months that the HHS data not only was pretty good, but it kept getting better
But we realized after a couple months that the HHS data not only was pretty good, but it kept getting better
Some criticisms turned out to be ill-founded, too. Many have sneered about HHS using Palantir, but HHS Protect was a kind of spin out of CDC's own Palantir system DCIPHER Cloud. And CDC's hospital data system is supported by a $60 million Leidos contract.
When we began looking more seriously at the HHS data system, a whole bunch of our data folks contributed to analyses of it. Big thank yous to @anthropoco, @cat_pollack, @NotoriousRSG, @betsyladyzhets @karaschechtman, and others for all their research.
If you want to look at the hospital data yourself, we now show it on the @COVID19Tracking state pages. @anthropoco and Kevin Miller also built this excellent map for weekly average facility-level data:
https://covidtracking.com/data/hospital-facilities
https://covidtracking.com/data/hospital-facilities
Here's the upshot of this story. A team of civil servants worked across a whole range of government agencies to make something really important. They did it quickly. They made sure the public could see the work. They've constantly improved.
And yet, some CDC officials (always anonymously) say the hospital data collection should move back inside the CDC.
I get that! In the abstract (and in the long-term), I can see why CDC is the best place for this data.
But 120k+ people are in the hospital right now!
I get that! In the abstract (and in the long-term), I can see why CDC is the best place for this data.
But 120k+ people are in the hospital right now!
To break the hospital data system would be an enormous gamble. The federal government's aid for hospitals is dependent on this data. The public's best view of hospital strain is dependent on this data.
And, to be totally honest with you, the CDC has moved too slowly time and again. Sure, the Trump administration tied their hands at times, but any honest reckoning has to look deeper into the institution. We need the CDC at its best.
Oh, and if you're looking for just the juiciest bit. That big hospital data changeover that supposedly shocked people at the CDC?
We saw correspondence that the CDC agreed to it after their own system couldn't add an important new data field for 3 weeks.
We saw correspondence that the CDC agreed to it after their own system couldn't add an important new data field for 3 weeks.
