vrdiver wrote:For a while now, we've been aware that the link between cases and deaths has been broken, or at least changing too fast to make historic numbers useful predictors of future trends. Many of us have started to rely on hospitalisation data as a better indicator of how this pandemic is progressing, at least in the UK.
I just read an article in The Atlantic* that points out that hospitalisation data is just as fraught with interpretation issues, as covid becomes more prevalent amongst the general population, more and more patients admitted to hospital will test positive for the virus, regardless of the reason for their admission. The article goes on to suggest that a better measure might be to track patients on oxygen, or with low blood oxygen levels.
The reason for my post is to raise a concern that, as winter approaches and the NHS workload increases for all the usual reasons, I fear we are going to see policy made based on covid prevalence in patients, regardless of symptoms.
Saying the NHS is overwhelmed, but not because of Covid19 may be an inconvenient truth too far?
VRD
*
https://www.theatlantic.com/health/archive/2021/09/covid-hospitalization-numbers-can-be-misleading/620062/
If you look at the graphs of
cases,
hospitalization and
deaths it is clear that the link between numbers of cases and hospitalization has also been broken. Albeit by factors of 2 rather than the factors of 10 for deaths/cases.
Deaths and hospitalizations are not simply proxies for the progress of the pandemic. They are measures of pain and suffering, things we care about.* Deaths and hospitalizations are a large part of the pandemic. If we are not talking about deaths or hospitalizations it's probably a lesser problem.
Deaths will always be an important metric. Lower deaths are good, not because they are proxy but because fewer people are dying.
Hospitalization is, as you say, an imperfect measure of suffering. Hospitalization, unlike death, is always optional (for somebody, not necessarily the patient) and standards change over time. Hospitalization are also an imperfect measure of efforts to control covid. There may be better metrics for both of these, but hospitalizations remains important.
When deaths and hospitalizations become low and stay low the worst of the pandemic is over. When history is written they are the numbers that will distinguish covid-19 and flu 1919 from Ebola 1994 or Flu 1976.
Policy makers are not historians, they should look forward rather than backward. Policy makers should not be using deaths and hospitalizations as a proxy for the disease. They should be using current and past deaths, hospitalizations, case, (and anything else they can get their hands on) as proxies for future deaths and hospitalizations. When the link between
current and
future conditions is broken in the right direction it's largely over.
* A sufficiently devious virus might arrange high levels of pain and suffering without deaths or hospitalizations. We would need other metrics. I cannot see how that would be as bad, but that may be simply be a willful suppression of imagination to avoid catatonia.