2 years in, we don't even have the numbers of people hospitalized from covid or with covid. It's not science reportings that is dead, it's science itself.

Currently about half of COVID-19 patients in hospitals are incidental cases who were admitted for another reason and then tested positive during admission screening.

https://gothamist.com/news/new-preliminary-state-data-shows-...

There are many people who have these numbers and study them.

Public health officials in the US have both the reported numbers as well as piggy backing on the system set up for influenza monitoring (CLI & ILI).

Simply because not all public reporting is as rich as, say, Illinois or Pennsylvania's, doesn't mean these figures aren't carefully scrutinized -- especially when threatened by political leaning from either side of aisle (such as when the then-current administration pressured the CDC to change and alter reports).

In my experience, hospital operations officers, health information exchanges, and, depending on the state, state HHS agencies don't mess around.

[0] https://www.cnbc.com/2021/04/09/trump-officials-bragged-abou...

Why don't they write/draw a Sankey diagram with something like 1 million people on December 1st 2019, with age bracket percentages, and then draw how some got COVID, some got hospitalized, some died, some got the vaccine, some got a booster, some died from things that are not COVID, all in proportion to how it happened in US population. The width of the flows would make it clear how effective vaccines are.

Sankey doesn't work well because we moved from SIR to SEIR+S with Delta -- meaning that we had breakthrough infections and the potential for waning immunity. Omnicron continues this.

[0] https://en.wikipedia.org/wiki/Compartmental_models_in_epidem...

Trying to draw a Sankey diagram that accurately describes what is happening are quite difficult: many of the numbers for those < 65 years old, and especially for those < 50 years old) are so small as to be hard to read: telling 1/1k vs 1/100k in a Sankey diagram doesn't work. They're both 1 pixel wide.