Long time lurker, rare poster here.. but I work in healthcare IT and specifically on COVID-related data and reporting (SQL and Power BI), so I feel I have something relevant to add. Now, I realize that my post is merely anecdotal and limited to a single healthcare system in the US, but I've also had interactions with people questioning the validity of hospital beds and ICUs being full, even within my own family and about the organization I work for, so I've got some feelings and words.

In an attempt to keep this brief (ha!

), some bullet points:
- CDC data is entirely dependent upon a long chain that begins with individual healthcare organizations, hospitals, clinics, and private practices, followed by county health departments, state health authorities, ending with the CDC itself. At various step along the way are thousands of people collecting, documenting, coding, aggregating, validating, and reporting the data.
- Individual healthcare organizations must abide by metric guidelines and reporting deadlines set by their local and state health authorities. These theoretically should be given to health authorities from the CDC, but there is (some understandable, some not) friction between the CDC and some states.
- An example of a metric guideline in my state: COVID-related deaths must have had a positive test within 60 days of the death date, otherwise the death is not counted as 'COVID-related.' I believe this matches the current CDC guideline.
- In the past 3 weeks, our local hospital has been over 90% capacity with our other nearby hospitals over 80%. We have over 300 beds total across 5 facilities. Roughly 40% of those occupied beds were COVID-related hospitalizations. These numbers are trending downward as of this week, and we hope that this trend continues.
- I, nor anyone I work with have ever received a directive to inflate, alter, or obfuscate our COVID numbers in any way. I work with both peers and physicians to ensure our data is validated and accurate before any reporting or submission. It's not always perfect, for example, we had to throw out about 70 tests at the beginning of the pandemic because they didn't have a location documented (someone made an error when the test type was created that wasn't immediately caught, so it flowed downstream for the 70 tests ordered under this type).
- The idea that there could be some grand conspiracy doesn't hold water for me. There are too many people involved in the data to feasibly deceive everyone and too much money on the line for hospitals to leave beds empty while claiming they're full.
One more point I'd like to illustrate before I go, and that is the way some states report their data to their county health departments, state health authorities, and eventually the CDC. Let's look at Colorado for the past 3 months as an example:
https://imgur.com/a/Tp67xLG (source:
https://usafacts.org/visualizations/coronavirus-covid-19-spread-map/state/colorado)
Notice that numbers are reported every day of the week except one (presumably Sunday) until after July 4th, then numbers are reported every day except weekends and holidays. You can even see the 7 day average line drop a bit for Labor Day, but the bars and lines seem mostly consistent over time and there aren't any large gaps or wild fluctuations in the data. They seem to be submitting their data consistently.
Now look at Florida for the same past 3 months:
https://imgur.com/a/znM2siv (source:
https://usafacts.org/visualizations/coronavirus-covid-19-spread-map/state/florida)
Look at June 19th to 25th, no cases reported. Look at July 16th, suddenly 75,000+ new cases. And no new cases reported for the following week? The bars and lines have gaps and are all over the place. Because that's how they have been submitting their data (or usafacts.org is part of a cabal that wants to make FL look inconsistent at data submission

). Unfortunately, it doesn't paint a very accurate picture of how the pandemic is going over there. If I were asked to analyze this data, I'd say it couldn't be done in a very meaningful way.
Now, not to get too political, but Texas and Florida have governors that have made their positions on the pandemic pretty clear, and I think we're seeing that affect their data in some way. I'm not saying it's necessarily nefarious, it could just be they're more lax on reporting requirements and deadlines which explains the funky FL example above. But if people are claiming that numbers in TX and FL are down and hospitals aren't overflowing and doctors and the media are all just lying about it, I'd 100% ask to see the data and make my own conclusion based on what I see.
From what I'm seeing here, I'm not sure I'd take Florida at their word. I'd instead check with the people actually working in healthcare there.
Thank you for coming to my Ted Talk.
[edit] Some corrections, and if anyone has any questions, I'd be happy to do my best to answer them. [/edit]