I think that your estimate of 2.5-10% population loss is derived from (50% infected) * (10-20% “severely ill” from the JAMA article) * (50-100% of severely ill patients die with severe overcrowding). I think that the last 2 numbers are questionable. Worldometers says that 8% of current cases are “serious/critical”, which as I understand it is roughly the threshold for being in the ICU/requiring breathing machine in normal circumstances. To get a 10-20% case fatality rate, we would need most patients with serious cases to die (feasible given lack of breathing machines), as well as a few percent of non-critical patients.
Estimates I made seem to line up with a fatality rate of less than 10%. I expect around 70% of critical patients to die without critical equipment (ventilators/vasopressors), so as of writing my estimate is in the ballpark of the following three numbers:
Diamond Princess: (7 actual dead + (36 critical * 70%))/706 = 4.5%
World actual CFR adjusted with formula from here (3,119 dead) / (3,119 + 48,163 recovered) = 6%
Higher fatality rate for overcrowding pushes this up, but on the other hand I think there is huge under-reporting of minor cases, since most countries are not testing everyone who was exposed.
WHO Report source: 5% “needed artificial respiration” * 70% + (10% “needed O2″ but not artificial respiration) * (wild guess 30% of these patients die) = 6.5%
Basically I’m confused at the gap between these reference numbers and the 10-20% fatality rate that seems to be required for the upper end of that 2.5-10% global population loss estimate. Do you just expect overcrowding to cause much bigger problems than shortages of breathing machines and other such equipment?
Notably eg the Diamond Princess number (36) isn’t critical cases, but “serious/critical” cases. Do you expect 70% fatality without critical equipment for the entire “serious/critical” category, or just the “critical” sub-category?
Further, curious about if there are explicable reasons underlying your 70% estimate, cause I am trying to estimate this myself too. Notably, I’ve seen numbers from China that 50% of critically ill patients die *with* medical care.
The “serious/critical” category on worldometer seems to mean “intensive care”, or “critical” as defined in the WHO report. Singapore’s 6% “serious/critical” means ICU admission. China’s 8.5% is close to the 6.1% on the WHO report—see below. Italy’s 10% is also intensive care, though their criteria for ICU admission could be different. It’s possible that the Diamond Princess number uses different criteria; do you know that they do? Anyway, the WHO said:
13.1% have severe disease [...] and 6.1% are critical. [...]
Severe cases are defined as tachypnoea (≧30 breaths/ min) or oxygen saturation ≤93% at
rest, or PaO2/FIO2 <300 mmHg. Critical cases are defined as respiratory failure requiring mechanical ventilation, shock or other organ failure that requires intensive care. About a quarter of severe and critical cases [i.e. about the number of critical cases] require mechanical ventilation while the remaining 75% require only oxygen supplementation.
Most ICU patients need mechanical ventilation, close monitoring for administration of drugs like IV vasopressors (1:1 or 2:1 nurse:patient ratio), or other scarce resources. But the inference from that to 70% mortality in an organized, severely-overcrowded hospital is mostly medical intuition. I have significant uncertainty from lack of reference classes, and also don’t know how well hospitals will be organized.
I think that your estimate of 2.5-10% population loss is derived from (50% infected) * (10-20% “severely ill” from the JAMA article) * (50-100% of severely ill patients die with severe overcrowding). I think that the last 2 numbers are questionable. Worldometers says that 8% of current cases are “serious/critical”, which as I understand it is roughly the threshold for being in the ICU/requiring breathing machine in normal circumstances. To get a 10-20% case fatality rate, we would need most patients with serious cases to die (feasible given lack of breathing machines), as well as a few percent of non-critical patients.
Estimates I made seem to line up with a fatality rate of less than 10%. I expect around 70% of critical patients to die without critical equipment (ventilators/vasopressors), so as of writing my estimate is in the ballpark of the following three numbers:
Diamond Princess: (7 actual dead + (36 critical * 70%))/706 = 4.5%
World actual CFR adjusted with formula from here (3,119 dead) / (3,119 + 48,163 recovered) = 6%
Higher fatality rate for overcrowding pushes this up, but on the other hand I think there is huge under-reporting of minor cases, since most countries are not testing everyone who was exposed.
WHO Report source: 5% “needed artificial respiration” * 70% + (10% “needed O2″ but not artificial respiration) * (wild guess 30% of these patients die) = 6.5%
Basically I’m confused at the gap between these reference numbers and the 10-20% fatality rate that seems to be required for the upper end of that 2.5-10% global population loss estimate. Do you just expect overcrowding to cause much bigger problems than shortages of breathing machines and other such equipment?
Notably eg the Diamond Princess number (36) isn’t critical cases, but “serious/critical” cases. Do you expect 70% fatality without critical equipment for the entire “serious/critical” category, or just the “critical” sub-category?
Further, curious about if there are explicable reasons underlying your 70% estimate, cause I am trying to estimate this myself too. Notably, I’ve seen numbers from China that 50% of critically ill patients die *with* medical care.
The “serious/critical” category on worldometer seems to mean “intensive care”, or “critical” as defined in the WHO report. Singapore’s 6% “serious/critical” means ICU admission. China’s 8.5% is close to the 6.1% on the WHO report—see below. Italy’s 10% is also intensive care, though their criteria for ICU admission could be different. It’s possible that the Diamond Princess number uses different criteria; do you know that they do? Anyway, the WHO said:
Most ICU patients need mechanical ventilation, close monitoring for administration of drugs like IV vasopressors (1:1 or 2:1 nurse:patient ratio), or other scarce resources. But the inference from that to 70% mortality in an organized, severely-overcrowded hospital is mostly medical intuition. I have significant uncertainty from lack of reference classes, and also don’t know how well hospitals will be organized.