If I am reading this correctly, the overall fatality ratio seems to be 0·657%.
Wednesday, April 01, 2020
Some science …
… Estimates of the severity of coronavirus disease 2019: a model-based analysis - The Lancet Infectious Diseases.
Subscribe to:
Post Comments (Atom)
Frank, the case fatality rate (CFR), defined as the percentage of deaths among people who test positive for the virus, is estimated in this paper to be 1.38%, almost 14 times higher than the 0.1% average CFR for seasonal flu. The 0.657% number is an attempt to estimate what they call the infection fatality rate, which would include those who had the virus but were never tested. To do this, the authors look at test data from six planeloads of people expatriated from Wuhan as the epidemic began to accelerate. Each passenger on these flights was tested, giving a small (extremely thin, as statisticians would say) data set for use in estimating the level of asymptomatic infected cases. Roughly half of the passengers who tested positive had symptoms so mild that they would never have been tested in China, and would not have been counted in determining the China CFR. On the basis of this data, the authors assume that around 50% of infected Chinese were never tested, and reduce the CFR of 1.38% to an estimated infection fatality rate of 0.657%.
ReplyDeleteThe layman needs to remember that like must be compared with like in statistics. The CFR for COVID-19 in this data set is 1.38%, while the CFR for seasonal flu is 0.1%. The infection fatality rate for COVID-19 in this data set is 0.657%, while that for seasonal flu is around 0.033% (two-thirds of flu cases are never reported). So, COVID-19 is fourteen times more deadly than seasonal flu when CFR's are compared, and nineteen times more deadly when infection fatality rates are compared.
Another like-to-like comparison for consideration. The first U.S. death from COVID-19 was recorded on February 29, a little more than four weeks ago. We now stand at 5,116 deaths.
ReplyDeleteAt an equivalent point in the 2019-20 seasonal flu season, we had recorded 56 deaths attributed directly to influenza, and no increase in pneumonia deaths from any cause. Think about it -- during the 2019-20 flu season we've gone from 56 cumulative deaths in Week 4 to nearly 35,000 cumulative deaths in Week 30. This is what epidemiologists mean when they talk about exponential growth. With over 5,000 cumulative deaths from COVID-19 in Week 4, where will we be in Week 30?
Anyone who thinks this is no more serious than a bad flu season is a fool.
According to the CDC, the estimated deaths from flu in this country between Oct. 1, 2019 and March 31, 2020 is between 24,000 and 62,000. That's a six-month period. The 5,116 deaths from Covid-16 multiplied by 6 is 30,696. That's more than the low end of of the CDC flu death estimate, but only half of the high end.
ReplyDeleteNow I admit to being more or less illiterate when it comes to statistics. But I have pretty good grasp of arithmetic. And we are talking death here, and every life is precious. But there are things worse than death, which at any rate is inevitable. I just wonder if we are not over-reacting and that commonsense precautions would be at least nearly as effective. I am 78 years old, which is to say in the age group most at risk for dying should I contract Covid-19. But I am also in very good health (and not just for someone my age, as a doctor has told me). I hope to hang around for a few more years, but if the price were ongoing quarantine, maybe not. But that's just me. Everyone else is free to be as scared as they wish.
Your grasp of arithmetic is not as strong as you seem to think. By simply multiplying the current number of deaths by six, you are treating viral transmission as a linear phenomenon. Wrong! It's an exponential phenomenon. Linear growth increases by a constant amount; exponential growth by a constant ratio. You should have learned that distinction in the eighth grade. I did. Perhaps the teachers at St. Ignatius Prep were better than those at St. Joe's.
ReplyDeleteYou, of course, are free to put your own life at risk as you see fit. But you have an ethical obligation not to harm others. I too must go out to buy groceries and meet other obligations (including shopping for several friends who are at greater risk than I), but in minimizing my time out, I am minimizing the chance that I might unknowingly spread the virus to others. With this virus, people are contagious for 2-3 days before becoming symptomatic, capable of spreading the virus before they know that they are sick. Equally important, I am lowering the chance that I will occupy an ICU bed or require a ventilator that should be available for workers at higher risk: first responders, grocers, doctors and nurses, etc. As the kids say: "It's not all about you".
Nobody is talking about an endless quarantine. The point is to prevent the health care system from collapsing under a surge of critical cases. We need time for hospitals to gear up for the onslaught: ICU capacity, ventilators, testing kits, personal protective equipment for doctors and nurses, medications, etc. Hospitals simply do not maintain large stocks of these things in normal times, and it will take a couple of months to put in place what's required. (My daughter is a hospital physician). The quarantine also buys time to investigate potential treatments. There will be no vaccine for at least 18 months, and no disease-specific treatment for 5-6 years. (My wife was a pharmaceutical scientist for decades). The only immediate route forward is to re-purpose existing drugs like remdesivir or hydroxychloroquine. But, again, large quantities of these meds are not simply sitting in warehouses ready to be used. They have to be made. Production plants have to be re-configured, raw materials must be purchased, manufacturing protocols have to be put in place, workers trained in the specifics -- on and on. Again, at least a couple of months.
A quarantine of a several months seems like common sense to me.
Actually, my arithmetic was fine. 5,116 times 6 is 30,696. The exponential business, I believe, is algebraic. And my college algebra teacher would agree with you, which is why he advised me to drop out of his course (which I did). My question is this: Have we seen exponential growth in this case yet? You believe that the number of deaths from Covid-19 in this country will increase exponentially. I am skeptical of that, given that there are already signs that it is leveling off. I hope — as I'm sure you do — that I am right, because fewer people will die. But it makes little difference in one sense. If I am right, the explanation will be that it was the draconian measures taken that kept the death rate down. So it's a no-lose gambit for the pessimists. But if I am right, should we not shut down everything once again come the flu season in order to minimize the fatalities that will inevitably occur during that? Should we not focus society on minimizing deaths in every way under all circumstances that we can forever and ever? If the numbers between flu and Covid-19 prove to be comparable, I don't see why that would not be the logical conclusion.
ReplyDeleteThe exponential growth in deaths is a reality, not a prediction. During the 24 hours since my second comment, 867 Americans have died from COVID-19 (data from Johns Hopkins Coronavirus Resource Center). In contrast, the initial 867 deaths in this country occurred over a period of 26 days. By definition, exponential growth involves not just an increase in some quantity, but an increase in the RATE of accumulation of that quantity. Exactly what we see here. It's got nothing to do with optimism or pessimism -- just realism.
DeleteAnd, yes, I truly hope that the death rate will level off, for everyone's sake, but especially because my daughter is a young resident physician at a hospital in a large metropolitan area. But reality trumps hope.
Have you no thoughts regarding your ethical obligation to avoid harming others? Or the value in giving our health care system some time to gear up?
Some thoughts about our toleration of 3,000-4,000 monthly deaths caused by automobile accidents, and another 3,000 to 4,000 monthly deaths caused by flu:
In the case of auto accidents, we seem to have made a complex, implicit socioeconomic decision based on how much we're willing to pay for automobile safety features, how much in taxes we're willing to pay for guardrails, traffic lights, etc., speed limits we're willing to tolerate, sleep requirements for truck drivers, and many other factors. Different countries make different socioeconomic decisions about what they consider acceptable in this area.
In the case of case of flu deaths, I suspect that the explanation is fairly simple: out of sight, out of mind. Most of these deaths occur among the elderly, the sick, and the poor. I know that physicians fight like hell for those patients, but the resources we as a society are willing to commit to the fight are limited.
On the other hand, people are incredibly fearful about meeting death at the hand of a stranger, and willing to pay enormous sums of money to forestall the possibility. Yet only 90 people a month meet that fate in the U.S. Can you imagine the panic in the streets if 3,000-4,000 people each month were killed by strangers?
We also spend a lot of money to research relatively rare diseases. Why? Multiple sclerosis, one of the better known such diseases, kills only 80 people a month in this country.
Trying to make moral sense of these disparities is pointless. Consequently, using any particular death rate as a guide for how to respond to COVID-19 is also pointless.
Here are some interesting numbers: https://pjmedia.com/instapundit/363586/
ReplyDelete