Best visualization I've seen of how age impacts COVID-19 outcomes, from data on 16749 patients.
Main takehome: people above 50 have elevated risk. Below that, less but nonzero risk. Other factors like gender or obesity are impactful, but age dominates.
medrxiv.org/content/10.110…
Main takehome: people above 50 have elevated risk. Below that, less but nonzero risk. Other factors like gender or obesity are impactful, but age dominates.
medrxiv.org/content/10.110…
The hazard ratio chart below is the crucial one. It shows which factors elevate or reduce risk of dying relative to the baseline of patients under 50.
Being female gives protection, obesity does increase risk, but age dominates.
So, let folks under 50 out?
Being female gives protection, obesity does increase risk, but age dominates.
So, let folks under 50 out?
As circled area on the left below shows, and @eladgil showed last week, people under 50 tend not to die.
But as the circled area on the right shows, a significant number do get seriously ill enough to be hospitalized, even if discharged alive.
But as the circled area on the right shows, a significant number do get seriously ill enough to be hospitalized, even if discharged alive.
I still think the best route is that of New Zealand. And I don't think it's wise to downplay the severity of COVID-19.
But if you wanted a cautious opening strategy that takes the disease seriously, it'd involve mandatory masks & folks under 50.
But if you wanted a cautious opening strategy that takes the disease seriously, it'd involve mandatory masks & folks under 50.
Simple, clean rules might help if reopening: wear masks, avoid crowds, stay at home if under 50, etc.
With that said, let's be clear that it'll be challenging to partition society in this way. Many mixed age households. And >30% of the US is over 50.
kff.org/other/state-in…
With that said, let's be clear that it'll be challenging to partition society in this way. Many mixed age households. And >30% of the US is over 50.
kff.org/other/state-in…
Two things about this study:
1) Shows crucial importance of EMR. They used it to get stats across 166 hospitals in quasi-realtime.
2) This is the kind of preprint I'm generally most confident in, because it's usually fully reproducible if you trust the underlying data table.
1) Shows crucial importance of EMR. They used it to get stats across 166 hospitals in quasi-realtime.
2) This is the kind of preprint I'm generally most confident in, because it's usually fully reproducible if you trust the underlying data table.
In general, I think any reopening that occurs prior to zero new cases may cause the virus to spread. Even if people wear masks, avoid crowds, and keep under- and over-50s apart, it may only reduce the effective reproduction number somewhat rather than push it below 1.
We'll see.
We'll see.
The ideal strategy:
- Get daily incidence to zero
- Keep effective reproduction number below 1
Meaning: first get to zero daily new cases, then inhibit the spread of the virus.
Anything less than this means the virus will likely eventually spread to a large number of people.
- Get daily incidence to zero
- Keep effective reproduction number below 1
Meaning: first get to zero daily new cases, then inhibit the spread of the virus.
Anything less than this means the virus will likely eventually spread to a large number of people.
How would you achieve this?
Well, NZ and Asian countries are close.
In the US it may not be feasible given state incompetence at all levels, and the nontrivial number of people who think it's just the flu.
But maybe you could close down borders at state or even county level.
Well, NZ and Asian countries are close.
In the US it may not be feasible given state incompetence at all levels, and the nontrivial number of people who think it's just the flu.
But maybe you could close down borders at state or even county level.
Put another way, you would first stop all travel into and out of your county, then turn it into a green zone without the virus, then gradually radiate out from there to link up with other green zones.
This requires competence as well. But on a local rather than national scale.
This requires competence as well. But on a local rather than national scale.
