3 ms·

Calculating “death rate” with a denominator of “recovered people” is simply lying with data: of all of the covid metrics, “recovered people” is the least well r

by irq11 6y ago

Calculating “death rate” with a denominator of “recovered people” is simply lying with data: of all of the covid metrics, “recovered people” is the least well reported. Of the millions and millions of people who have now had this virus and recovered without incident, only a small fraction have seen the inside of a medical facility. Only a tiny fraction of those have had “recovery” reported, and even when it is reported, it generally lags by months. The only reason you would divide “deaths” by that number is to mislead people and scare them. The overall IFR for this virus is somewhere around 0.1-0.6%. That is the actual chance a person will die if they catch it. For healthy people under the age of ~50, divide that by ten (at least).

ddingus 6y ago
Given the deaths in my peer group of known, tested, infected people, the higher number I cited makes more sense. We shall see as more outcomes continue to be known. We will get better at tracking that because this virus is more likely to be endemic, forced. Having some healthy respect for this mess, rooted in basic fear and concern for the future, is just not a bad thing. As we get better at the data related to chronic complications, that info will also give people a more solid reason to live and work differently. Fact is a whole lot more people than we think will suffer a reduction in overall health too. Also not a bad thing. Frankly, I believe the outcome based death rates will fall over time as we improve our will and ability to record outcomes. I do not believe it will end up being fractions of a percent. I also believe complications, organ damage, etc... will be respectable percentages as we improve on capturing that data. We are finding immunity does not endure. We are also finding repeat cases can vary widely in severity. Add in the growing reality of it being endemic, and there is a solid basis for real concern. Ideally, we all remain able to revisit all this after a full year cycle.
iso1210 6y ago
> The overall IFR for this virus is somewhere around 0.1-0.6% Do you have a good citation for this, because the Biostatics research university at Cambridge have IFR 10 times that. https://www.mrc-bsu.cam.ac.uk/now-casting/ https://www.mrc-bsu.cam.ac.uk/now-casting/
- irq11 6y ago
  That’s just ridiculously, inexcusably wrong. The best estimates have been under 0.7% since March. The CDC current estimate is 0.65%: https://www.cdc.gov/coronavirus/2019-ncov/hcp/planning-scenarios.html https://www.cdc.gov/coronavirus/2019-ncov/hcp/planning-scena...
  - iso1210 6y ago
    So the CDC say 0.7% as the best estimate, above even your highest 0.6% statement. Until recently Cambridge had 0.9%, this has increased recently as more evidence has come out. Taking a 0.7% value would mean that 40% of Brooklyn has had it, a figure not backed by serological tests. Even is 100% of Brooklyn had has covid, that would give and IFR of nearly 0.3%
ddingus 6y ago
Let's coin toss 100 times. Let's put 'em in orbit and assume they survive reentry and land where we can evaluate heads or tails, so we have some flight time in this analogy. Just for shits and giggles. It takes a coin anywhere from two weeks to a month to land. We have tossed all 100 coins, 25 per week over the course of a month. Our stats could look like this. Tossed coins = 100 Landed coins = 50 Tail landings = 4 To make it a bit more interesting, we could also say we are tossing 25 new coins per week right now, and that could change up, or down. New coin tosses this week = 25 Right now, we know the outcome for 50 coins, and 4 of them landed tails. (4/50) = 8 percent tails rate We know at least 50 coins are in flight, outcome unknown. It could be that every other coin toss lands heads. We could also say, there are no more tosses to consider. So let us add the 25 new tosses this week too. (4/125) = 3.2 percent tails rate. This would be the best case. It could also be every other coin lands tails too. (54/100) = 54 percent! That is the worst possible case. In reality, for this example, we know there are more tosses coming and we also could model poor testing and say we are not aware of all the tosses. Some error factor could be computed to refine best and worst case too. The better we are at collecting data, the more those outcomes will tell us. The more time that passes, the more outcomes will tell us. We may find tosses at night double risk of landing tails, for example. Or maybe coins tossed mid week land tails consistently. Ignoring that which has not come to conclusion biases the risk assessment toward lower risk with a larger error. Low numbers with high error rates are not all that useful in this context. Current global stats are: Coronavirus Cases: 19,725,865 Deaths: 727,548 Recovered: 12,660,022 That's 5.7 percent deaths overall known outcomes with 7,065,865 cases in progress right now. If people test positive, don't end up in the hospital and we have no death certs filed, they are recovered, but may be harmed in some way. Poor data on that. Otherwise, we know they died. Some argue a percentage of those deaths are with COVID causing a death due to other factors. You can bias that overall 5.7 percent number up or down based on your own personal risk assessment, but dead is still dead, and catching the virus only adds to ones risk. These all count for purposes of policy and understanding danger. I know at least one person who did that low risk assessment, went to a State with high case load and who has been on a ventilator for a month now. They may live, and that looks increasingly likely, but will definitely not be the same after it is all said and done. "I am unlikely to die" carries greater weight in that light, yes? All I am really saying here is people would be wise to treat this with more respect than we may otherwise feel is warranted. The basic reasons are simple ignorance (we are doing the science and it takes time) and lack of effective medicine at this time. Both of those warrant the more firm, lower error numbers we get from actual outcomes.

by irq11 6y ago

ddingus 6y ago

Given the deaths in my peer group of known, tested, infected people, the higher number I cited makes more sense. We shall see as more outcomes continue to be known. We will get better at tracking that because this virus is more likely to be endemic, forced. Having some healthy respect for this mess, rooted in basic fear and concern for the future, is just not a bad thing. As we get better at the data related to chronic complications, that info will also give people a more solid reason to live and work differently. Fact is a whole lot more people than we think will suffer a reduction in overall health too. Also not a bad thing. Frankly, I believe the outcome based death rates will fall over time as we improve our will and ability to record outcomes. I do not believe it will end up being fractions of a percent. I also believe complications, organ damage, etc... will be respectable percentages as we improve on capturing that data. We are finding immunity does not endure. We are also finding repeat cases can vary widely in severity. Add in the growing reality of it being endemic, and there is a solid basis for real concern. Ideally, we all remain able to revisit all this after a full year cycle.

iso1210 6y ago

> The overall IFR for this virus is somewhere around 0.1-0.6% Do you have a good citation for this, because the Biostatics research university at Cambridge have IFR 10 times that. https://www.mrc-bsu.cam.ac.uk/now-casting/ https://www.mrc-bsu.cam.ac.uk/now-casting/

irq11 6y ago
That’s just ridiculously, inexcusably wrong. The best estimates have been under 0.7% since March. The CDC current estimate is 0.65%: https://www.cdc.gov/coronavirus/2019-ncov/hcp/planning-scenarios.html https://www.cdc.gov/coronavirus/2019-ncov/hcp/planning-scena...
- iso1210 6y ago
  So the CDC say 0.7% as the best estimate, above even your highest 0.6% statement. Until recently Cambridge had 0.9%, this has increased recently as more evidence has come out. Taking a 0.7% value would mean that 40% of Brooklyn has had it, a figure not backed by serological tests. Even is 100% of Brooklyn had has covid, that would give and IFR of nearly 0.3%

ddingus 6y ago

Let's coin toss 100 times. Let's put 'em in orbit and assume they survive reentry and land where we can evaluate heads or tails, so we have some flight time in this analogy. Just for shits and giggles. It takes a coin anywhere from two weeks to a month to land. We have tossed all 100 coins, 25 per week over the course of a month. Our stats could look like this. Tossed coins = 100 Landed coins = 50 Tail landings = 4 To make it a bit more interesting, we could also say we are tossing 25 new coins per week right now, and that could change up, or down. New coin tosses this week = 25 Right now, we know the outcome for 50 coins, and 4 of them landed tails. (4/50) = 8 percent tails rate We know at least 50 coins are in flight, outcome unknown. It could be that every other coin toss lands heads. We could also say, there are no more tosses to consider. So let us add the 25 new tosses this week too. (4/125) = 3.2 percent tails rate. This would be the best case. It could also be every other coin lands tails too. (54/100) = 54 percent! That is the worst possible case. In reality, for this example, we know there are more tosses coming and we also could model poor testing and say we are not aware of all the tosses. Some error factor could be computed to refine best and worst case too. The better we are at collecting data, the more those outcomes will tell us. The more time that passes, the more outcomes will tell us. We may find tosses at night double risk of landing tails, for example. Or maybe coins tossed mid week land tails consistently. Ignoring that which has not come to conclusion biases the risk assessment toward lower risk with a larger error. Low numbers with high error rates are not all that useful in this context. Current global stats are: Coronavirus Cases: 19,725,865 Deaths: 727,548 Recovered: 12,660,022 That's 5.7 percent deaths overall known outcomes with 7,065,865 cases in progress right now. If people test positive, don't end up in the hospital and we have no death certs filed, they are recovered, but may be harmed in some way. Poor data on that. Otherwise, we know they died. Some argue a percentage of those deaths are with COVID causing a death due to other factors. You can bias that overall 5.7 percent number up or down based on your own personal risk assessment, but dead is still dead, and catching the virus only adds to ones risk. These all count for purposes of policy and understanding danger. I know at least one person who did that low risk assessment, went to a State with high case load and who has been on a ventilator for a month now. They may live, and that looks increasingly likely, but will definitely not be the same after it is all said and done. "I am unlikely to die" carries greater weight in that light, yes? All I am really saying here is people would be wise to treat this with more respect than we may otherwise feel is warranted. The basic reasons are simple ignorance (we are doing the science and it takes time) and lack of effective medicine at this time. Both of those warrant the more firm, lower error numbers we get from actual outcomes.