...
Remember please that the percentages Oren cites are from reputable literature, BUT that study was based on a 2.3% CFR for the population. We know that the CFR for the hospitalized population is double that. So - double ALL of the numbers in the table.
...
For the US, I suspect that the most likely death toll will be about 5 million. I cannot see it being less than 1 million. The high end is likely 25 million.
The high end comes about from presuming that there is not some large unseen portion of the population that never shows anything other than mild symptoms and is never tested; plus presuming the 4.5% CFR is accurate and representative for the hospitalized portion (who then represent everyone); and that half of those requiring intensive care die as a result of the collapse of the hospital systems when they become over run. That last part adds about 7.5% to the CFR of 4.5%, hence a 12% death rate. Italy is already seeing close to that. And they aren't through the worst of it yet.
...
Sam
Sam, why do you assume that we have a 4.5% CFR in the USA? Would appreciate if you could explain that! Latest data show a much lower figure.
I believe that too many analyses are focused on the natural mortality rate with treatment, rather than on the number whose symptoms are severe enough to require hospitalization which is a far higher percentage than occurs with the flu. Even with the flu affecting a higher count each winter than COVID-19 has thus far, the impact on the hospital system is already far outpacing the flu.
The problem with many analyses and forecasts is that they don't take into consideration the issue of overwhelming of the intensive care units (ICU) in the healthcare system, as seen in Italy and Spain, and as will SURELY be seen in the USA, France and UK in the coming weeks. About 6% of the infected in C19 will need ICU, which is way more than for the regular flus.
In a way, the situation is on one hand that we wish for everything to happen quickly, get to herd immunity in the population (=more than 90% infected for this virus), and then C19 will be just like any other flu virus, as we will not have a vaccine or mass vaccination in the nearest 18 months or so. From this point of view, we want the infection to run freely so we can become immune.
On the other hand, we fear death, we don't want people to die. So we want to slow the infection down, and above all suppress it sufficiently so that the ICU units aren't overwhelmed. If 6% of your cases require intensive care and you can’t provide it, most of those people die. As simple as that. We don't want Italy or Spain.
Meanwhile, the disease continues to progress exponentially in many big countries, e.g. USA. Inflexion point on the positive exponential growth has still not been established. Meanwhile, efforts to measure and contain it seem to be more linear.
Korea has been very actively testing and tracking infected people. Korea tracing/testing has shown that around 11% of positives proceeded to Serious (=supplemental oxygen) or Critical (=ventilator). About 1/3 of those infected in S/C state will eventually die. That is in a health care system that is NOT overwhelmed. We have much worse ratios in overwhelmed Italy.
I think it's not the best metric to use total cases, because ‘active’ cases have yet to be determined. Now that we know the disease course—5 day mean incubation, 9-10 days symptoms, then either recovery or serious/critical on day 11-12, the best CFR proxy is Fatalities/ recoveries. We have a 3.0% CFR ratio in
S.Korea of resolved cases, and we have only 1 % S/C out of active (unresolved) cases.
Moreover, of active unresolved cases, 6.0 % in
Italy are S/C, implying that about 45% of those S/C die and 55% eventually recover in a medical system that is overwhelmed.
These data are all readily available on
https://www.worldometers.info/coronavirus/country/About USA, we're already getting reports of overwhelmed hospitals/ICU units in some states. And we still have a couple of weeks of exponential growth left. There are 4 million admissions to the ICU in the US every year, and 500k (~13%) of them die. Without ICU beds, that share would likely go closer to 60-70%. Even if only 50% died, in a year-long epidemic you go from 500k deaths a year to 2M, so you’re adding 1.5M deaths, just with collateral damage.
So, on one extreme we have countries like S.Korea, on the other extreme we have the overwhelmed countries. I think any forecast for USA has to take into consideration the issue of the forthcoming overwhelming of ICU units.