Millhouse

Our ability to answer that question is increasing by the day.

And in there hides the basic answer in all of this: media coverage and interest

If you really want to avoid COVID move to Vanuatu - apparently one of the last places to not have it. Be careful of the locals though as they can get a bit bitey.

scrubbedexpat141

Yes and no, you could argue it's being answered by the day but the answer is just as negative as a lockdown situation isn't it?

The basic answer is those countries (and many others) simply don't care about those issues enough to do anything. They're not on the doorstep and affecting them. It's not pleasant but it's a reality.

Millhouse

One interest stat (which undermines my own arguments don't get we wrong) is that around 1600 people die per day in the UK - that's the baseline. Covid is adding around 700 to that right now. Which on the face of it is alarming, but the reality is that - as you say - many of those 700 would be in the 1600 anyway. Services will be stretched but it's not the total wipeout we are being told it is.

Oh, and if I get it - as an asthmatic that can bearly process a cold, I'd probably be brown bread.

csdf

We don't shut down the country over heart disease because there are cheaper, less dramatically impactful ways of materially reducing heart disease deaths. Similarly, we don't ban cars because there are cheaper, less dramatically impactful ways of materially reducing car crash deaths. Unfortunately, without a vaccine, or evidence of successful ways to treat the illness, the only current way of materially reducing covid19 deaths is through social isolation.

Miss Ann Thrope

If you are going to look at deaths, then the relevant measure in terms of capacity to cope is the daily rate. Covid-19 daily deaths are already far in excess of that from car crashes in most of the affected countries viz. average UK daily car crash deaths is five and we're now above 400 for Covid. You really think that supports your point? In fact people in car crashes now are more likely to die because the hospitals and emergency services are so overloaded.

Car accidents don't also incapacitate 20% of the productive population due to illness and quarantine (the very conservative modelling of what would happen with no controls).

Covid will exceed 1m dead with all of the controls in place. Without them, they would likely have far exceeded car crashes and potentially all other sources of death combined. It is a specious argument now of course, but if you were in power, would you be willing to take that gamble?

Malaria is an interesting one because most of the deaths are in poor countries. The failure to develop an anti-malaria vaccine is a clear indictment of the way we have privatised most major pharmaceutical research. Big pharma isn't pursuing it because there is no money in it. The governments of the countries most affected don't have enough money left to do the research after their senior officials have filled their Swiss bank vaults. I agree it's a scandal but it doesn't invalidate the case for controls to prevent the spread of Covid-19, one of the most contagious epidemics ever to emerge (and this is before any mutant variations have had a chance to develop).

The Covid controls are to AVOID armageddon. Some countries acted early and avoided such dramatic action. Some thought they had and seem to have got it wrong...Japan, oft quoted by the sceptics seems to be getting into a mess as we speak.

Miss Ann Thrope

Brilliantly put: this is the point.

Miss Ann Thrope

Without controls, health services will be swamped so that hundreds of thousands - likely millions - more people globally will die, of all sorts of things, than would have otherwise died. And a huge proportion of them would be the healthcare workers that we need to look after us. Look at what has happened even where there are controls. Have you not seen any of the footage from hospitals in Lombardy?

scrubbedexpat141

"Yeah but 500,000 people a year die from flu"

ETC.

weasel decentral

Yeah no major dramas, after the initial period of panic most of the food is readily available in the shops or by delivery. There's a lot of overblown media stuff focusing on India, police heavy handedness etc. but by and large everybody is reasonably compliant.

scrubbedexpat141

The videos of the rozzers whacking everyone with sticks are eye-opening for sure.

Deliveries here are starting to struggle. A french supermarket have had my cash for 14 days and delivered squat. Part of me is just assuming they've surrendered completely.

FriendlyExpat

It would be a huge error to believe the modelling is conservative. All the data sets being used by epidemiologists are wildly inaccurate. Oxford University is doing a regularly updated forest plot of death rates and it demonstrates the outcome from different countries is so completely disparate that you shouldn't rely on them. They won't even do a point estimate. This is the core problem - garbage data in, garbage models out.

https://www.cebm.net/covid-19/global...atality-rates/

Miss Ann Thrope

As we all know (presumably), modelling an epidemic relies on a set of assumptions much like any other mathematical model. Most useful models will allow for a range of assumptions to be tested, for example the progression of the R0 value as the epidemic in any given territory progresses. The 20% simultaneous population incapacitation rate with no controls is at the low end of outcomes predicted by most reliable models (principally IHMES but also Imperial College when they did their famous adjustment which led to the abrupt change in policy by the UK government). So I think it's fair to characterise that as conservative.

On the other hand, it's a pretty wilfull misreading of my comment to infer that I think all the modelling driving public policy is conservative because I don't. But it is clear that there has been a great deal of wishful thinking in the early stages of many government responses, choosing to dismiss bad news projections because of unreliable and contradictory data. Similarly in this forum. But that just shows that they didn't grasp how epidemiological modelling works. Uncertainty about the scale of many factors does not mean we can't determine a few big important things and acting early always reduces the range and scale of bad outcomes. The key is to choose the right actions of course. Dithering and then finally capitulating (Spain, UK, USA) is rarely the best option. Though data is evolving rapidly and of variable quality, it is just plain ignorant to dismiss the usefulness of expert (I use the word very deliberately) models. Well designed models, which can still provide meaningful output even with high degrees of uncertainty around the input, are the most powerful and useful drivers of public policy during an epidemic.

The IHMES models are probably the most widely accepted globally and have the distinct benefit of evolving to reflect the latest data and information. They are currently predicting over 60k UK deaths by August by the way, even with all the restrictions that have been put in place. The burning issue in the coming weeks in the UK will be the critical shortage of all hospital beds (and possibly staff) and in particular ICU places (less than 10% of projected requirements). That may be a big contributor to the high rate of UK deaths in the model as that situation is much worse here than elsewhere, even worse than in Italy.

So it's not just about fatality rates, which even the linked article acknowledges are of limited use (because the denominator is so completely unknown without mass testing), and massively variable. Other important factors can be reasonably characterised based on relevant knowledge from past epidemics. The longer we go on with Covid-19, clearly the more we know. While the models will thus improve, they will naturally become less useful as the policy dies will have already been cast. But the incontravertible evidence from Italy and Spain makes abundantly clear that the priority must remain urgently stopping too many people getting sick too soon. Otherwise we will have a far higher price to pay than that currently being exacted.

FriendlyExpat

IHMES? What about JHU, the WHO, the many university / epidemiological teams and other national health bodies looking at this?

As for conservative, Imperial is using a point estimate with a 95% confidence interval. If you think that's conservative...

Also note Imperial assumed a 0.95% IFR; Oxford University today predicts 0.1% to 0.39%, and they caution that is likely still a overestimate due to data collection methodological flaws see: https://www.cebm.net/covid-19/global...atality-rates/

Miss Ann Thrope

Again, I am not saying any MODEL is conservative. The confidence interval will be for the described outcome based on the assumptions used. Most decent models, including all those you mention, can be run with a range of assumptions to determine a range of possible outcomes. A 95% confidence interval gives you that level of confidence in the outcome, if the assumptions are valid, no more than that. The assumptions are what is critical. That's the essence of modelling. So those models can be run with sets of assumptions which are conservative, say with an R0* less than 2.5, or aggressive, say with an R0 of 6 or above. Each of these scenarios appears to have been the case of how this epidemic progressed in some places and at some times. The purpose of the models is to determine the likely outcomes against any one of these assumptions (and a bunch of others, often inter-dependent). Policy is determined on the basis of which assumptions are considered most likely and which outcomes must be favoured or completely excluded.

My point, again, is that even conservative assumptions indicated unacceptable outcomes in most of the models, without the imposition of social controls. Hence why almost every affected country has imposed some sort of controls (even the oft cited Sweden). And the evidence suggests these controls are having the desired impact even though most were implemented very late and thus some countries have had to bear a lot of pain before that impact is felt.

Fatality rate, at this stage in the epidemic, is slightly academic. It's very difficult to establish yet with any level of certainty because it depends on really knowing the total number of infection cases, and as you point out, that number is often very unreliable or incomplete. IFR is not the factor driving immediate policy decisions right now and uncertainty in IFR is not yet a valid basis to challenge the policies that are being implemented. Total fatalities, number of critically ill and number of people incapacitated are much more important outcomes to understand at this stage and these can be measured with high degrees of precision. Ultimately IFR is critical to understand to manage this disease going forward and hence why these data are being accumulated. At the moment those data are just too uncertain and incomplete to be useful.

[*Note: R0 is the number of further infections generated by any one case. Normal policy objective is to drive it, and keep it, below 1 to stop the epidemic. Without a vaccine or cure, the only mechanisms available to do this are social controls and physical barriers such as PPE.]

FriendlyExpat

Which models are you talking about? Imperial (in the UK) was alarmist. Most others were not as alarmist (although there was a wide range).

Countries are like sheep; once one country imposes a restriction, politicians in other countries are forced to as well otherwise the electorate in the non-compliant country believe they are more at risk. A race to the bottom ensues where the most restrictive restrictions spread worldwide, even if they make no sense. Think school closures / border closures / flight bans etc.

If you think the lockdown has empirically worked, explain Japan? No lock down, 4,200 cases, 93 deaths vs UK - heavy lockdown, 61,000 cases, 7,000 deaths? Singapore & Sweden also.

https://coronavirus.jhu.edu/map.html

This is not correct. There is a wide variance in the way countries record fatalties (e.g. Italy - dying with Covid counts as a covid death, in UK - dying OF covid counts as a covid death). And Oxford says no significant increase in overall UK death rate can be seen: https://www.cebm.net/covid-19/england-and-wales-mortality-during-the-covid-19-outbreak

R0 is not the right measure - you need to be looking at Rt. Also note R0 is not a "truth", it's merely an empirical measure that varies depending on numerous external factors.