In reply to Offwidth:
> I don't know how many times I have to be clear what my opinion is and why.... the most important covid response data IMHO is now NHS general data. I do want to get the message across, so will post elsewhere as well (if new NHS threads start).
Today's dashboard update is uncomfortable on first glance, but it needs some careful thought. I've reached beyond healthcare data I'm familiar with, so I hope one of several familiar faces can chime in to confirm or correct me!
The rate of admissions is hovering around constant in most regions now - the rate constants have stopped plunging for negatives and are levelling off. This isn't the behaviour we'd expect following the cases curve, nor is it the behaviour that's really needed.
Now, it's worth thinking about the numbers a bit.
Thinking about the behaviour of cases and the forwards projected uncertainty shading from yesterday's ONS update [1] we probably have about 7% of people infected with Covid now. In a recent pre-Covid year, there were 17.1 million "admissions episodes"in the year to English hospitals [2], or about 47,000 per day (ignoring seasonality).
If we assume I've found a reasonable source for hospital admissions and that all admissions are PCR tested (I await correction if I'm wrong)....Take those pre-covid daily admissions and give 7% of them incidental Covid, that gets us to ~3,300 incidental covid admissions per day.
Now, that's interesting because we're seeing ~2,000 covid admissions a day in England.
I expect there's some anti-correlation between having high-immunity omicron infection and going to hospital, as infection has spiked harder in younger people and older people are more vulnerable to all-cause hospitalisation.
So, perhaps we don't expect to see much more decrease in the admissions data on the dashboard as we move to endemic circulation and incidental admissions?
This also ties with the primary data LSRH cited above and linked to further plots thereof, which shows that in several regions including London, about half of all Covid admissions are now incidental admissions (patients are not primarily being treated for Covid)
Evas as things move towards" with Covid" not "of Covid", the problem remains of people in hospital having Covid, and of this requiring more isolation measures (for now), but as Covid itself puts fewer and fewer people in to hospitals, suppressing its spread by population-wide NPIs to help with infection control in hospitals becomes a thorny issue, because such suppression starts re-charging the pandemic potential in the population; it seems to me that more sustainable solutions are required, and given that there are many other respiratory diseases it seems that better control measures against nosocomial transmission are a key part of this; I hadn't appreciated until a discussion with minimike on another thread that some Blair era PFI hospitals are basically air-tight boxes and potentially far worse than older parts of the estate in terms of spread of Covid.
Two other thoughts the data and estimates on incidental admissions raise:
- Incidental deaths - harder to bound I think. Death in 2019 were ordered of 530,000, which with a 7% Covid prevalence applied uniformly across the population gives incidental deaths (with Covid) of ~100/day. It's only the subset of these that occur in hospitals that I would expect to be detected as Covid cases. Again anti-correlation may emerge from the demographics of infection.
- The next big question becomes what the endemic prevalence of Covid will be, and how much that will phase-lock to seasonality in the weather over the next couple of years.
I think at this point, interpreting the dashboard data is rapidly teetering towards being a fool's errand. Well, for England at any rate. Poster maggot shared a link on this subject the other day on another thread [4]
[1] https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/con...
[2] https://digital.nhs.uk/data-and-information/publications/statistical/hospit...
[3] https://www.ons.gov.uk/aboutus/transparencyandgovernance/freedomofinformati...
[4] https://www.independent.co.uk/news/uk/uk-end-covid-pandemic-b1991297.html
Post edited at 17:06