Latest thread covering COVID-19 admissions and deaths in English hospitals, with an updated R estimate.
While the PHE dashboard hasn’t yet been updated, @NHSEngland data indicates 870 admissions today. https://t.co/xkI8zMNbmu
— COVID-19 Actuaries Response Group (@COVID19actuary) October 20, 2020
Based on recent hospital admissions our estimate of R remains at around 1.3 at the last point we can estimate.
Obviously there is significant regional variation in both current admissions and current growth rate, which we discuss below. 2/4 pic.twitter.com/EPfKD2QX3Q
— stuart mcdonald (@ActuaryByDay) October 20, 2020
Average COVID-19 deaths in English hospitals is now over 80 a day, and it looks like Saturday 17 October will be the first day with over 100 hospital deaths in the second wave (once recording is up to date).
The doubling time is around ten days based on the recent trend. 3/4 pic.twitter.com/neH3BDoSSD
— stuart mcdonald (@ActuaryByDay) October 20, 2020
Regional situation here👇 https://t.co/4yn5LifyhL
— stuart mcdonald (@ActuaryByDay) October 20, 2020
And here’s the admissions data on a log scale (we’ve had a lot of requests for it). https://t.co/x0NQc2v4MH
— stuart mcdonald (@ActuaryByDay) October 20, 2020
Every admission (and even day elective surgery) requires a test, and has done since very early on, so community testing levels will have no impact on admission numbers. (For elective surgery if the test is positive, it doesn’t happen.)
— John Roberts (@john_actuary) October 20, 2020
Thanks for this. I guess the question is how many of these admissions will be displaced by covid if any? Considering flu didn’t materialise in the Southern Hemisphere and how that impacts capacity. Amazing how different flu seasons can be too.
— Ryan (@1nationtory) October 21, 2020
Looking at charts of current Covid trajectory relative to that in March April suggests a much less frightening picture. Why do trends only show the most recent weeks, when a proper comparison of deaths (not infection rates) and numbers admitted to hospital or ICU are most relevant to assess the situation more broadly? Genuine question. I am really puzzled that statistical charts are focussing only on the recent past which distorts the longer term perspective.
Isn’t it interesting how we’ve stopped talking about Sweden. I assume the totalitarians are concerned that a different approach that’s allows people to make their own choices, works…
Hi Ros
The COVID actuaries produce weekly bulletins which provide bigger picture stories and have done a large number of bulletins, everything they have produced is available at http://www.covid-arg.com .
They are not lobbying for any particular position and I try to feature all bulletins and twitter threads on this site – again without adopting a particular position.
I’m always nervous about comparisons with the first wave, which often come with a suggestion, implied or otherwise, that unless and until we reach those levels, everything is OK. As the families of 50,000+ lives who contribute to excess deaths statistics will testify, the first wave was far from OK, and we should be striving to avoid getting anywhere near those levels. More so, as the only way the health service was able to cope with the first wave was by cancelling much routine work, which has its own medium and long term health implications.
Thus our focus is thus on the second wave, and how rapidly it is developing, which our graphs show. Hospital admissions are currently growing consistently at around 40% per week, and for those wanting a comparison with the first wave, it’s easily possible to see how quickly some regions may match the first wave.
It’s clear that the gradient of growth is much less severe, as we would expect with a much better understanding of how to protect ourselves. That’s a really positive place to be in, and buys us some extra time to avoid the scenario we saw in April. But unless we get R below 1, growth will continue, and we will reach those levels, albeit in a matter of weeks, not the days we saw in late March.
Outcomes are better too now, although the mortality rate is still too high for those entering hospital. But I don’t think we should underestimate the long term implications on the health of those who emerge from ICU units alive. And Long COVID appears to affect those who had a less severe bout of illness too – the health affects of this virus go way beyond the number of daily deaths reported.
In conclusion, we’re looking at the dynamics of the current position, not looking over our shoulder to compare against the appalling situation we found ourselves in in the Spring. I don’t think such comparisons are helpful, other than to try and ensure we do much better this time.
Can you please explain where the quantity “Estimated Occurences” comes from in the graph of hospital deaths? On 17 October they have jumped to 25% of reported deaths.
They are an estimate using a common actuarial technique used in insurance claims based on observed reporting delays. The most recent days will always have the highest “add-ons”. The graph omits the last two days as we have found that there is considerable volatility in reporting to make them less reliable. For interest, the total for the 17th rose a further 10 to 106 today. Hope this is helpful.