Ricochet is the best place on the internet to discuss the issues of the day, either through commenting on posts or writing your own for our active and dynamic community in a fully moderated environment. In addition, the Ricochet Audio Network offers over 50 original podcasts with new episodes released every day.
More Fevered Calculations: Working the Coronavirus Numbers
In all the hype and happy talk around the latest coronavirus to cross over to humans, keep an eye on this number in America: 498,000. That is the number of people this novel coronavirus will have to infect to cause as many deaths as the annual, seasonal flu. I tried to make sense of the numbers around notorious coronavirus, a.k.a. COVID-19, in a post about a week ago now.
I now note that the presidential proclamation, suspending travel from certain countries, referred to COVID-19 as “SARS-CoV-2.” The CDC page explains the reason for the changing names. This prompted another look at the numbers, with this math-challenged scribbler doing a bit of stubby pencil, back-of-the-envelope figuring. Check my math as I work through the numbers; hopefully it is better than Ma and Pa Kettle’s.
SARS last time had an 9-10% mortality rate. It was nothing like influenza, including the worse known version in 1918-19, because the number of people who actually contracted SARS was so much lower than the number of people who get the flu globally every year. Think about it: if the flu regularly kills 1 in 1000 infected, and SARS killed 1 in 10, then you can see that SARS would have to infect 1/100 the number of flu victims to reach the same death count.
It is still very early days, perhaps 90 days into this latest coronavirus outbreak, but the deaths divided by total resolved (recovered or dead) percentage (likely the best measurement of lethality) is down to 6%, according to the Worldometer coronavirus dashboard. So, the question will be how many cases are detected that recover versus die over the next few months or year. If 1 in 1000 die of flu, then I think that means this new coronavirus variant would have to infect 1.66 percent as many people as the flu to kill the same number this year.
[It is fair to note that the numbers out of China and Iran may be unreliable. We should probably track deaths divided by total resolved cases excluding these two large sources. However, the early numbers from each reporting country vary wildly, suggesting a single number, capturing all data from all sources outside China and Iran, may itself be misleading in these first 90 days.]
While 1.66 percent sounds small, the sheer number of flu cases is always huge. How huge? That is hard to say, because it is only an estimate and varies widely year to year, according to the CDC:
CDC estimates that influenza has resulted in between 9 million – 45 million illnesses, between 140,000 – 810,000 hospitalizations and between 12,000 – 61,000 deaths annually since 2010.
That is just in the United States. Eyeballing the numbers, it looks like the median number is around 30 million falling ill with the flu annually. So, 0.1% mortality means around 30,000 deaths expected from the flu in such a year. If my fevered figuring is right, and if SARS-CoV-2 (COVID-19) only has to infect 1.66% of that flu number (30 million) to extract the same death toll, then this novel coronavirus will have to infect 498,000 people in the United States to cause as many deaths as the annual, seasonal flu. Sure enough, if you multiple 498,000 cases by 6% mortality rate, you get roughly 30,000 deaths.
This is why we should all be interested in limiting the spread of this latest bad bug. Let’s all cough and sneeze into our elbow, not our hands, wash our hands while singing the alphabet song, and work on getting good sleep, food and a bit of exercise for good health. Oh, and check the CDC COVID-19 page from time to time. I leave you with three videos for the weekend: Ma and Pa Kettle, the alphabet song, and Fever.
First up, math gone horribly wrong:
Second up is the original (and only legit) alphabet song, brought back to my attention in a tweet by a woman to the effect of “Washing my hands when a little girl started singing the alphabet song. Couldn’t stop until she finished!”
Finally, let’s all strive to avoid giving each other the wrong sort of fever:
[Thanks for the feedback, leading to a correction on SARS, a clarification on the mortality metric, and a brief note on data quality.]
Published in Group Writing
This post is part of our Group Writing Series under the March 2020 Group Writing Theme: “Working.” There are plenty of open days, so get busy and work it! Stop by and sign up now.
Interested in Group Writing topics that came before? See the handy compendium of monthly themes. Check out links in the Group Writing Group. You can also join the group to get a notification when a new monthly theme is posted.
I’m nervous about this – and comparing it to the flu. Reason being are all the unknowns. We have a flu vaccine not nothing for this virus. It has been spreading so fast, and containment seems to be difficult, even here in the US. I don’t remember anything like this. SARS and other outbreaks didn’t cause schools to close, events canceling, religious sites closed, flying test kits to a cruise ship that can’t dock, and so many more things. Also, while I realize Trump is pushing for them to accelerate getting a vaccine, does that include testing and determining side effects? Banks and the Feds are taking emergency measures, airlines are saying it has a 9/11 feel? Also when the gov. tells you not to panic, isn’t it time to panic? I’m not saying I’m panicking…………gulp……..Spring break just started here – the place will be swarming with kids and families. Also, not sure what info to trust. Every day another Iranian official is going down….more cases in Italy. Gulp…
I don’t the like the data used. The number infected we have is based on people presenting to a doctor and being diagnosed. That might be throwing out 90% of infected that don’t go to the doctor. Until there is randomized scientific testing, it would be better to compare number-deaths over number-hospitalizations for flu verse number-deaths over number-recovered for Covid-19. That is about 8% vs 5.6%. But those numbers are dominated by China, which we can’t trust.
Depends on who’s telling you not to panic. But it’s true that the government’s broken trust with the public (a problem they created themselves) cannot be easily repaired.
Also, remember that the 2009 H1N1 epidemic lasted a full year, with an estimated 60 million cases in the U.S. and more than 12,000 deaths in the U.S.
How this will compare is anyone’s guess, but I don’t remember a great deal of panic back then. Why not? What’s changed?
The bigger you think the number of infections, the less lethal this is. I agree we should stick to numbers outside of China, with the deaths being fairly reported and the number of cases being assumed low end.
If you assume as I do that most cases go undiagnosed, then the mortality rate might be much, much lower. Making the net result much like the flu.
No. Do the numbers. Do not buy the political hype, nor the political happy talk. The way you keep the total number of cases down is by common sense measures to limit exposure and transmission. The flu shot is a best guess, a cocktail of around three of the strains estimated to be the most likely to spread in a given year. You can get a flu shot and then come down with the flu because you got a strain of the bug not included in your vaccine. Yes, it improves your odds.
Watch the number of deaths and reported cases, tracking number reported recovered in the short term. Does the U.S. number of deaths look anything like it could possibly be 30,000 in this year? We are three months into it. People flew back and forth between China and the U.S. for two months before any travel restrictions. If there are actually many more people here who were exposed, who caught the bug, then the lethality is much lower than current estimates.
The party in the White House, of course.
I still can’t discern if there is merit to all these precautions or if this is a case of public hysteria that makes global warming fears seem tame.
Just in the past few days, I’ve seen the mortality rate estimated anywhere from 1% to 10%. I’m not sure which sources are most trustworthy. And now there are reports of multiple strains.
This will likely sort out over the rest of the year, with accuracy of reporting on this disease approaching reporting on the flu.
Math. You get the low end number if you compare all estimated cases to reported deaths. You get the high end by something like the number of people cured versus dead. People are talking past each other.
You’re so cynical.
I approve.
When you consider all resolved cases (recovered or dead) then it’s between 5–6%.
What about those who “recovered” and got sick again, like the person in San Antonio? Are there enough cases like that to skew the stats?
Early days, still small “N,” but “recovered” and then sick again means an adjustment to “recovered.”
This counts the cases we know about. Low-grade cases that were never reported/tested or asymptomatic cases are not included.
Regular flu viruses also have mild cases which go under-reported. Estimates of unreported cases shouldn’t affect the overall statistics, which are probably comparable.
And those two numbers will converge over time. In the short run, because the number of unresolved cases are high, the death rate calculated as deaths/(Deaths+recovered) is high. Given enough time, that calculation will be equal to deaths/cases. For mainland China the first death rate was 5.5% as of yesterday, that is down from 7.7% a week earlier. The deaths divided by cases has been between 3.5% and 3.75% over the same period.
Right. Which would make that figure even lower.
The things that concern me about nCov 2019 are its long incubation period, and reports of re-infection. The incubation period means it’s already everywhere so it’s something we’ll have to manage. The virus that caused SARS may have been deadlier once you got it, but it was relatively contained and wasn’t as readily contracted as this new one.
nCov 2019 is an RNA virus, so it’s going to be hard to pin down. Smallpox was a killer, but at least DNA viruses don’t mutate like influenza – that’s one reason we were able to eliminate it. If those reports of second infections in the same individual are correct, that implies this little bugger mutates quickly.
So, we have a newly widespread, and easily spread, virus that can cause severe pneumonia in susceptible individuals. It may be expected to present a moving target to the vaccine manufacturers.
Deaths / (deaths + recoveries) is the best measurement, isn’t it?
I have my grave doubts about the numbers in Iran. Last I checked using the Johns Hopkins site–admittedly some time ago–the death rates were dramatically different in Hubei, mainland China outside of Hubei, and the rest of the world. The best explanation seemed to be that some deaths in mainland China outside of Hubei were not being reported.
These days when I want to check on the death rate I just run the deaths / (deaths + recoveries) equation through the Worldometers numbers for all locations except mainland China and Iran.
I just got about an 18% death rate.
Now deaths happen faster than recoveries, so there’s a glimmer of hope. But I am not an optimist about an estimate around 6%.
I don’t follow this math.
As I’ve pointed out before, the percentage of deaths basically has no meaning. The denominator is the people with symptoms who have been tested and are positive for the virus. So this leaves out people who are infected who do not have symptoms, people with symptoms who have not been tested, and false negatives. How many are in each group? We don’t know and have no way of knowing. (There are also factors affecting the numerator, but I’m getting deep in the weeds as it is.)
I work in the busiest ER in a metropolitan area with a population over one million in Texas. According to @rodin, from his latest post, there are 16 cases in the state. How many patients with symptoms have been tested at my ER? Zero.
Don’t regular flu statistics have the same problem?
Yes, but it’s a lot easier to test for the flu. For the flu, we just do a single nasal swab and put it in a machine; the results are available in less than twenty minutes. For the Corona virus, we have to take three samples: one from each side of the nose, and a different type of swab from the mouth. The samples are then sent to another facility for processing, and don’t come back for 24-48 hours. I do flu tests almost every day. I have never done a Corona test.
Good to know!
It’s wrong. The correct number is 1/100, which is even scarier.
To get one flu death requires 1000 cases. To get one SARS death requires 10 cases. 1000/10 = 100.
An argument based on math should always be checked and rechecked.
Is this a common epidemiological statistic? Seems odd to me.
Deaths are obvious, “recovered” requires a clear illness with convalescence. A better measure of lethality would seem to be deaths/contracted, with “contracted” to include those who do and do not show an illness. Tough data to obtain, no doubt.
Do we have an MPH or an epidemiologist among us?
In my Inbox today, from the Mass Board of Registration in Medicine (emphasis added):
“Dear Physician Colleagues,
Massachusetts continues to prepare for the potential outbreak of the novel coronavirus, COVID-19, that originated in Wuhan, Hubei, China. The risk of COVID-19 in the Commonwealth is low; in contrast, the risk of flu is high.
Our Department of Public Health staff have been in constant contact with the Centers for Disease Control and Prevention (CDC) and state and local partners, and launched a site that includes extensive information and is a “source of truth” at: http://www.mass.gov/2019coronavirus. The site includes resources and fact sheets translated in multiple languages that can be printed and shared.
Please continue to practice yourselves and share with your patients that these simple precautions will help prevent the spread of the flu and other respiratory illnesses:
• Cover your mouth when you cough or sneeze, using a tissue or the inside of your elbow
• Wash your hands for 20 seconds with soap and warm water frequently and use hand sanitizer
• If you have a fever or feel sick, stay home and call your healthcare provider
• Get the flu vaccine – it is not too late! Find more information on the flu and where to get a flu vaccine here: https://www.mass.gov/info-details/what-should-i-know-about-flu
To date, there has been one confirmed case of Coronavirus in Massachusetts. As of March 6, the Massachusetts State Public Health Laboratory has confirmed seven presumptive positive cases that are awaiting confirmation by the CDC. However, there have been more than 30,000 confirmed cases of the flu. The risk for COVID-19 in the Commonwealth remains low; the risk for the flu is high.
If you have clinical questions about testing or a patient in your care then please call the Epidemiology Line at 617-983-6800; the Department of Public Health epidemiologists are available 24/7 to support you.
Please stay informed by visiting both the MDPH and CDC websites. To prepare for potential community spread of COVID-19, please think about your own personal preparedness and visit our preparedness website to learn more about developing a preparedness plan for you and your family.”
To me, this says it all.