3 ms·
The fact they haven't published their models worries me. At a minimum, I'd like to know what their models are optimizing for. Is it "Minimum number of deaths
by londons_explore 7y ago
The fact they haven't published their models worries me.
At a minimum, I'd like to know what their models are optimizing for. Is it "Minimum number of deaths" or "Minimum economic cost", or "Minimum economic cost taking into account a cost of $XM per death".
Minimum number of deaths probably isn't a smart metric, because the solution to that is to tell 99% of people to stay home in bed for the rest of their lives. Hard to catch a disease if you never leave bed.
Minimum economic cost (with a factor for cost of a death) is hopefully the model they've used, but it's politically suicide to accept that grandpas life was 'worth' $X.
- sgt101 7y agoIt's pretty challenging to get a model published in the time span required - I believe that the team at Imperial College are racing to do this, but given the behaviour of the virus is still a topic of research it's unsurprising that they are not out there at this time.
- londons_explore 7y agoA model can begin as a spreadsheet with a handful of guesstimates in. You can start making decisions based on such a model. Then in parallel, you get teams refining the model and making it more accurate, while other teams work on implementing those preliminary decisions, potentially changing course if newer, more refined, models come up with different results. There's no reason the crude model can't be published 1 hour after it was made. If anything, publishing the model will help localized branches of government make their own decisions ('hmm, looks like we'll be expecting lots of ambulance use at the end of march, that isn't a good time for the fleet inspection!'). Private companies would also benefit from the model, and can better schedule logistics. Even a hugely inaccurate model is better than "who knows! panic!".
- tigershark 7y agoA model has nothing to do with the underlying data. If they thought that the model was good enough to justify the deaths of hundreds of thousands of people they should release it right now, so we all can take a look at that abomination and decide for ourselves if they should be prosecuted for mass murder in a couple of months.
- sgt101 7y agoIt's here now : have a look. https://www.imperial.ac.uk/media/imperial-college/medicine/sph/ide/gida-fellowships/Imperial-College-COVID19-NPI-modelling-16-03-2020.pdf https://www.imperial.ac.uk/media/imperial-college/medicine/s...
- mytailorisrich 7y agoIs that the one that seemingly a U turn under the pretence of "new data"? [1] Because the government already knew that rapid herd immunity would cause a huge number of deaths. [1] https://www.theguardian.com/world/2020/mar/16/new-data-new-policy-why-uks-coronavirus-strategy-has-changed https://www.theguardian.com/world/2020/mar/16/new-data-new-p...
- sgt101 7y agoIt is new data - the assumptions that were in the model showed that the surge capacity of ICU could cope in a social distancing scenario, but once they factored in the rates from Lombardy they got a different result. The horrible truth is that this thing is going to cause a huge number of deaths, I wish it weren't so.
- grey-area 7y agoThe health services do make difficult choices like that every day though. NICE (National Institute for Health and Care Excellence) restricts expensive medicines which can save lives for that reason - to spend the money on something else. Rationing happens and it will be necc. in the coming months in every country, whether that is by patient income or by outcome, I’d prefer it by outcome personally.
- tigershark 7y agoPlease, try to tell this to uk general public when pretty much 1 people over 2 or 3 has one of their parents killed by the Coronavirus because of boris inaction. I think what will happen to you will not be a pretty sight.