UPDATE March 19th: A lot can change in a few weeks. Most of the case statistic used in this article turned out to be very much skewed by the poor state of testing availability, something that we are still struggling with. It still remains to be seen how weather will impact the spread–the main question being whether or not the added temperature + humidity + testing capacity will drop the R0 below the point of exponential spread. Right now we cannot assume that that will be the case, so our primary focus should be on limiting spread via social distancing, especially with regards to the elderly, and on doing what we can to boost our own immunity. For my full immune-boosting protocol and up to date info on the latest science behind the virus, check out my new project: covid19immune.com
Some other considerations: the current mortality rate is still being consistently overestimated, as the poor state of testing means that, of those who are exposed, generally only the ones who experience severe symptoms get tested and therefore counted as “cases.” It is likely that the total Case Fatality Rate will end up being no higher than the 0.6% out of china (6x the flu), and likely lower than even that. Another promising factor is that there are multiple antiviral drugs that have finished some early trials showing very promising results for treating the disease (and will hopefully further lower the mortality rate and health system burden).
For those still interested in the original article, read on, but keep in mind the updated situation.
TL;DR – Based on current data, the lethality of the coronavirus for someone who is young, healthy, and has access to quality medical care is ~0.018%. The odds of getting it in the U.S. in the next month is ~0.015%. Total odds of dying from it: roughly 1 in 36 million. The flu is less talked about but a far bigger problem.NOTE: The high mortality rate for the sick and elderly means that we should still be attempting to make sure they are never exposed to it. Just as we should be doing with the flu.
Is the above number wrong? Almost definitely. It is based on the most recently reported statistics, which themselves are likely to be inaccurate due to lack of comprehensive testing and a low sample size. Some factors that could easily worsen the odds are:
- If the current number of cases in the U.S. has already put us significantly farther along the exponential curve.
- If our healthcare system becomes overburdened
- If you are over 65 or have a pre-existing medical condition.
Update – March 13th: The first point on this list ended up being the case. We are now at a point where it is a reasonable possibility that anyone might be exposed over the next few months. It is still unclear what affect the changing weather may have. Our priority should therefore now be to slow its spread (limiting the burden on the healthcare system) and to boost our own immune systems as much as possible. I will be publishing more resources on these topics soon.
There is no doubt that the coronavirus is a dangerous disease. The question is simply how we should approach thinking about it, and how we should let said thoughts impact our behavior. There is a difference between being cautious and being paranoid. In this case, the best way to begin grounding our thoughts in rationality is to look at the available facts and statistics and compare this disease with similar ones that we already have a better understanding of–keeping in mind that the comparison is apt but imperfect.
Last week, 2,714 people died in the U.S. suffering from severe respiratory disease. The previous week the number was 3,386, and in the second week of January it was a whopping 4,180. 1
None of these people had COVID-19 (previously known as “2019 novel coronavirus”).
Instead, they were suffering from the far more commonplace influenza, a.k.a the flu (and its more dangerous progression–pneumonia).
The flu comes every year. We’re used to it. It’s no longer new and scary–it’s just a fact of life. Reporting on it doesn’t generate advertising revenue.
The problem with modern media is that it causes us to blur the lines between voyeurism and practicality. Our brains evolved systems to ensure that any danger signals in our environment hold our attention and provoke immediate fight-or-flight reactions. This worked very well for most of human history, where we were only capable of seeing things in our immediate vicinity and said reactions were appropriate. Now, when we can pull out our phones and see people dying halfway across the world, this circuitry gets hijacked. The problem is further compounded by the availability bias, which makes us more likely to worry about dangers that we have recently seen or heard about regardless of how statistically likely they are to actually apply to us (e.g. seeing a news report of a plane crash makes us temporarily afraid to fly).
I find that the best way to combat this tendency (other than avoiding the news entirely) is to trust in the numbers.
There are two main questions we need to answer:
- How deadly is the coronavirus?
- How likely am I to contract the coronavirus?
In both cases, we’ll be comparing the new danger to its most-similar incumbent: the flu.
Coronavirus Vs. The Flu – Which is more deadly?
Let’s start by comparing the top-level statistics. We’re getting towards the end of the flu season (more on this later), so we’ll be comparing the projected flu season stats with the current precise coronavirus numbers. The goal of this is to get a sense of scale and corresponding relative risk. Note that this is a comparison between the global coronavirus numbers with the U.S.-only flu statistics. This is because so far there are only 88 reported cases in the U.S.–not enough to get any useful stats on the disease itself.
Based on these numbers, it would seem that the coronavirus is indeed significantly more deadly than the flu, displaying a case-fatality rate (CFR) of 3.4% vs 0.1%. But these statistics need to be taken with a large grain of salt. At this early stage of the outbreak, it is very difficult to tell precisely how many cases there actually are (especially because the majority still relies on data coming out of China). It’s far easier to report a death than to report a sickness, and the specific test for COVID-19 is still both expensive and not widely available. It is likely that the actual CFR for the disease is significantly lower than the above numbers reflect.
But should it be a global concern?
The flu kills an average of 389,000 globally each year.2
This means that as of right now, the coronavirus has caused 0.7% the number of deaths expected to result from the flu worldwide, which itself is only ~ 2% of all annual respiratory deaths. Unless it continues to spread in an exponential manner (more on this soon), it is unlikely that the coronavirus will have a significant global death-toll. See below for other causes of death that are even more impactful than the flu.
Now on to our continued calculations.
We know that the majority of the mortality risk for the coronavirus is taken on by the the oldest segment of the population. An analysis of over 70,000 cases from China shows the following breakdown:
It’s clear that by far the largest risk is for those over the age of 65. For those in the lower age groups, it was found that the patients who are at most risk are the ones who already have severe health conditions. From the study: “While patients who reported no comorbid conditions had a case fatality rate of 0.9%, patients with comorbid conditions had much higher rates—10.5% for those with cardiovascular disease, 7.3% for diabetes, 6.3% for chronic respiratory disease, 6.0% for hypertension, and 5.6% for cancer. Case fatality rate was also very high for cases categorized as critical at 49.0%.”
Reason #1 Not To Worry About the CoronavirusIf you have no existing health conditions and are under 65 years of age, you have less than a 1 in 1000 chance of dying if you do get sick.
Another significant factor is that this study was conducted on the outbreak in rural China. A study on the the global influenza mortality rate found that health care access, socio-demographic development indicators, and baseline respiratory mortality explained more than two-thirds of the variance in influenza-associated respiratory mortality rates between countries in people <65 years 4. It is likely that a similar statistic would hold in the case of COVID-19. Another analysis showed that the overall mortality rate for the coronavirus outbreak was significantly correlated with the healthcare burden5 (outcomes became worse as hospitals in China were overloaded). Especially for patients that develop pneumonia or other serious complications, access to proper medical care can be the determining factor in survival.
Reason #2 Not To Worry About the CoronavirusThe data show that if you live in an area with access to high quality healthcare, your mortality risk is significantly lower than that reported from the China data.
Now let’s look at the numbers for the flu.
It’s pretty clear that if you get the flu, it won’t be pleasant, but there is little actual risk as long as you are relatively young and healthy. NOTE: In order to make a better comparison with the coronavirus age-group demarcations, the data from this graph was pulled from a study that grouped deaths and transfers to hospice care as the negative outcomes. While this may seem significantly different, these transfers typically happen when the flu has caused further complications that, while not being instantly lethal, often have an extremely severe impact on quality of life and mortality risk.
And what if you are not young and healthy? 31.5% of all Americans, almost a third of the population, are now living with multiple chronic conditions (MCC), a number that jumps to 80% of people 65 and older.8 All of these people are at a significantly higher risk. In one study of hospitalized adults age 65 years or older, patients with laboratory-confirmed influenza were more likely to experience persistent catastrophic disability, shown by a significant decline in functional independence and activities of daily living one month after discharge.9
Among adults hospitalized due to flu, for whom information on underlying medical condition was available, 92 percent had at least one reported underlying medical condition that placed them at high risk for flu-related complications, the most frequent of which included
- cardiovascular disease (46%)
- metabolic disorders such as diabetes (43%)
- obesity (37%)
- chronic lung disease (30%)10
All of these same statistics apply in the case of the coronavirus, where they are similarly worrisome. The key point to note is not that there is no risk, but instead that these have always been significant risk factors, and our fear of them should be tempered by the relative risk of exposure.
It is also worth noting that for both the flu and the coronavirus, mortality rates for infants are extremely low. While our instinct is always to protect our children, their systems are often even more resilient than our own.
How does the coronavirus stack up vs the flu?The coronavirus is more deadly than the flu, but in either case this is only a risk on an individual level if you are either very old or already sick.
Will The Coronavirus Kill Me Or Those I Care About?
I am currently 30 years old and healthy. Based on age, that puts me in the 0.4% mortality risk category. I have no comorbid conditions, so as stated earlier that puts me at 5-10x less risk, as those cases cover the majority of deaths in this age bracket. I also live in New York, with access to high quality healthcare in the event of complications, which based on the earlier statistic should give an added 3x survival factor. I could add in another multiplier based on my health being above average for the population, but even without that we already have me sitting at a 0.018% mortality risk. That is IF I contract the disease. A similar calculation could be applied for most of my peers.
Now let’s run this same calculation for my parents age bracket, currently on the border between the 50-64 range and the 65-74 one. Assuming the higher 3.6% risk, and thankfully no comorbid conditions (in my parents’ case), we have a 0.16% mortality risk. This does start to get significantly higher as we get to the age of grandparents or start adding on other health concerns.
But again, this is all only if we actually get the disease.
Should I be worried about transmitting it to those who are more susceptible?Yes you should be. This analysis by no means claims that there is no risk, and if you are exposed to it then you should be careful about coming into contact with the elderly / infirm. It is still significantly more deadly than the flu, and the long incubation period is a cause for concern. People should be taking extra precautions to wash their hands, avoid coughing or sneezing on others, etc. Using humidifiers is especially important.
What Are The Odds Of My Contracting The Coronavirus?
Diseases are characterized by their R0 value–the average number of new infections that an infectious person can generate in a population that was not previously exposed to the virus. An analysis of current studies on the coronavirus found the average R0 to be 3.28 and median to be 2.79, which exceeded initial WHO estimates of 1.4 to 2.5 11.
If we look at the graph of current total coronavirus cases, the situation seems promising. This is largely due to the spread of cases in China seeming to have slowed (although again, it is unclear precisely how reliable those numbers are).
However, if we look at the cases outside of China a different story seems to present itself (hence the cause for the current state of global panic):
This graph looks scary. Until we adjust the zoom.
As of January 26th, there were 56 cases outside of the U.S. Over the course of the following month, that number shot up to the current 4,289 cases.
To put this number in perspective, between 1999 and 2014, there were 10,206 deaths in the United States from “Accidental suffocation and strangulation in bed.” 12
If we assume a similar exponential growth within the U.S. (i.e. a failure of quarantine), that would mean that in a month from now we’d be facing ~5,000 coronavirus cases. This would likely lead to ~160 deaths assuming the same mortality rate as in China, but it is likely that that number would be somewhat lower given the aforementioned factors.
The U.S. population is 327.2 million. Let’s assume (somewhat arbitrarily) that I, living in NYC, am 10 times more likely to be exposed to an infected person than someone living in rural America. This still leaves me with a 0.015% chance of contracting the disease.
Reason #3 Not To Worry About the CoronavirusBased on current projections, there is less than a 1 in 5000 chance of my contracting the disease.
Now that we have calculated both a mortality risk and an infection risk, we can multiply the two together: this gives me a 0.0000028% chance of dying from the coronavirus in the next month.
Comparatively, my odds of dying in a car crash this year are .01% 13.
Let’s put that in more friendly terms:Based on the known statistics for the coronavirus's lethality combined with its current rate of exponential growth in the United States, I have roughly a 1 in 36 million chance of dying from it in the next month. Click To Tweet
That’s less than half as likely as my being struck by lightning in the same time interval.14
Yes, all of these calculations make a lot of assumptions and are based on incomplete data. Given the low availability of coronavirus testing kits in the last few weeks, it is possible that the starting number for the U.S. cases might be significantly higher than listed here. But they don’t need to be 100% accurate. Even if they are off by several orders of magnitude, it should still illustrate why the current state of panic is overblown.
Why Do Your Calculations Only Extend One Month From Now?
The answer is simple: we are approaching the end of the flu season.
If we look at deaths from the flu over the course of the last 5 years, it becomes extremely apparent that they tend to be limited to the winter months. Once the spring hits, the numbers drop to almost zero. As seen from this graph, the season already started to drop off two weeks ago.
Why is that, and why should it apply to the coronavirus?
The spread of airborne diseases is impacted by both the ambient temperature and the humidity. In 2007, a researcher named Dr. Peter Palese measured the influenza transmission rates between guinea pigs kept in adjacent cages at varying temperatures and humidities. His results are shown here:
Another study found that simply adding a humidifier to classrooms reduced the incidence of influenza-like illness by a factor of 2.3 16. A literature review on the topic looked at both laboratory and epidemiological studies and showed that humidity was highly correlated with both transmission rates and survival17. The precise mechanisms for this are manyfold, but one component is that, in a more humid environment, dry skin particles carrying the virus are more likely to encounter a water droplet and fall to the floor rather than be transmitted. Similarly, any expelled water droplets in a cough or sneeze are more likely to acquire more moisture from the humid air. When this happens they become too heavy to remain airborne.
Other factors that play into the increased spread during the winter are:
- People spend more time indoors in close-quarters.
- Lowered immunocompetence due to factors such as lowered Vitamin D and melatonin levels.
- The coronavirus is surrounded by a lipid layer that is only stable at colder temperatures (it breaks down as ambient temperature begins to rise)
It is impossible to say with 100% confidence that the coronavirus will be just like the flu in its seasonality. However, given that all of the above listed factors are equally applicable to both illnesses, it seems likely that the same trend will hold. This could be one reason that the total active cases graph from above looks the way it does, as temperatures in China are already beginning to rise.
No, we don’t know that the coronavirus will be exactly the same as the flu in its seasonality, but given that the transmission mechanisms of the two are the same 18, the above social and immune factors are the same, and the humidify-based transmission inhibitions are more a matter of physics than of biology, it seems likely that the same trend will hold.
CONCLUSION:Even with the exponential growth rates that have been seen so far and a starting value of 88 U.S. cases, it is likely that the change of season will halt the spread of the coronavirus in the US before it becomes impactful..
Are Governments Wrong To Worry About The Coronavirus?
No they are not. Organizations such as the CDC and the WHO are doing exactly what they are supposed to be doing: tracking a new disease and doing what they can to limit its spread. Even if this only saves hundreds or thousands of lives, it is worth the effort. This same type of quarantining would not be possible with the flu because of how widespread it is from the start of the season.
It is also worth noting that even if the changing weather halts its spread in the northern hemisphere, it is possible that it will remain an issue in some southern countries and then still be a problem for us come next year. This is yet another reason why it makes sense for these organizations to figure out the best ways of mitigating its impact.
There is, however, a large difference between the government organizations doing their jobs and the current hysterical reactions of the general public, as spurred on by the media. Our brains are programmed to ignore statistics, and the news cycle takes advantage of this fact.
Whenever these situations arise, it behooves us to run the numbers and think: “Is this actually something to worry about enough to change my behavior?”
Put It In Perspective: Other Recurring Health Problems We should Be More Worried About
- Some 829,000 people are estimated to die each year from diarrhea as a result of unsafe drinking-water, sanitation, and hand hygiene 19
- 1,010,000 die each year from HIV.
- 102,000 die each year from other STD’s 20
- 446,000 die each year from malaria
- 793,000 die each year from suicide
This list is just the tip of the iceberg. These issues may not be new and exciting, but they still need our attention, and our funding, to help solve. For anyone who has been moved to donate money to help with the current coronavirus problems, I would urge you to perhaps divert some of those funds to something like the Against Malaria Foundation–consistently sitting at the top of the GiveWell list for charities where your money will have the greatest impact.
One Final Note
This analysis was conducted based on the currently available data. As far as I can tell, the numbers check out and my current stance on the issue is the appropriate one. As with any risk analysis, we do have to take into account the odds that what we don’t know will end up having a large impact, and in this case the transmission details of the virus are still poorly understood. If more data comes out that seems to substantially impact the analysis I will attempt to adjust accordingly.