- Joshua Konstantinos
Coronavirus - The Outlook Has Worsened
With better data the range of outcomes has narrowed and the likelihood this blowing over seems small.
Last week I wrote that there were not a lot of hard facts to go on when trying to estimate the likely outcome of the novel coronavirus outbreak. Firstly, because there was simply not a lot of data available, but also because as I said last week:
…the numbers from China are inherently unreliable (lets not forget that China deliberately drove sick SARS patients around in ambulances to hide them from the World Health Organization back in 2003)
In the past week the case against trusting China’s data has strengthened considerably. Barron’s reported that an analysis of the variance in China’s data showed it was too perfect to be real data:
A statistical analysis of China’s coronavirus casualty data shows a near-perfect prediction model that data analysts say isn’t likely to naturally occur, casting doubt over the reliability of the numbers being reported to the World Health Organization.
Additionally, CNBC reports that high placed administration officials have stated that the US does not have “high confidence” in China’s coronavirus figures.
So while the numbers coming out of China have seemingly improved, it is crucial to watch what the Chinese government does and to keep a careful eye on the data from outside mainland China.
Lockdowns in China
In China, more and more of the country is in quarantine. The New York Times reported that as of February 15th it was estimated that some 760 million people in China are under some form of lockdown. Check out the excellent map below (Source) which is the best visualization I’ve seen of the extent of the lockdowns in China.
Areas in red under some form of lockdown
Given the spread to countries around the globe, we can now make some important characterizations of the virus without relying on China’s data.
Case Fatality Ratio
One of the most important factors of an epidemic is determining the Case Fatality Ratio (CFR). We now have enough cases from outside mainland China that we can estimate the CFR independently — this has the added benefit that it includes a significant amount of asymptomatic cases which would be difficult to detect within China.
There are various statistical methods of calculating CFR during an ongoing epidemic. Here, I use the equation:
where, D(s), R(s), and C(s) denote the cumulative number of deaths, recoveries, and cases, respectively.
Why use this methodology? Firstly because of the unreliability of the data from mainland China adds uncertainty to other, more sophisticated, statistical approaches. Moreover, studies on calculating the CFR during the SARS epidemic showed that this approach was “reasonable at most points in the epidemic” although it briefly underestimated the true CFR at one point during the 2003 outbreak.
As of this writing there have been 5 deaths outside of mainland China and 135 recovered patients Source for recoveries and deaths. This gives us a CFR of 3.70% (3.40% — 4.01% CI: 95%)
Confidence Interval and Margin of Error calculated by:
n = [z2 * p * (1 — p) / e2] / [1 + (z2 * p * (1 — p) / (e2 * N))]
Where: n is the sample size, z is the z-score associated with a level of confidence, p is the sample proportion, expressed as a decimal, e is the margin of error, expressed as a decimal, N is the population size.
This is a fairly high CFR — particularly if you assume it will rise if healthcare systems become overwhelmed. It is notable that the official Chinese CFR is significantly higher in the hard hit Hubei province than officially reported in the rest of China..
I’ve also estimated an R0 value of the virus of 3.66. (What is a R0 Value?)
This estimate is the weighted average from all the published studies I’ve been able to find, with studies whose calculations based on newer data weighted slightly more.
The weighted average R0 value is the average R0 values of the studies below, with each study weighted to lose 2% for each day earlier than the latest published study. E.g. a study based on data from 1/22/20 would be weighted 30% less than a study with data up to February 5th (15 days later).
Preliminary estimation of the basic reproduction number of novel coronavirus (2019-nCoV) in China, from 2019 to 2020
Novel coronavirus 2019-nCoV: early estimation of epidemiological parameters and epidemic predictions
Reporting, Epidemic Growth, and Reproduction Numbers for the 2019 Novel Coronavirus (2019-nCoV) Epidemic
These estimates do not paint a good picture for the world.
It’s important to note that the CFR estimate is still based on a fairly small sample size, and that the R0 value is not a constant feature of the virus, but a measure of infectiousness that can be lowered with containment measures. That being said, despite China’s draconian containment measures, the virus has continued to spread outside of mainland China.
Moreover, while the virus is not as fatal as some of the other outbreaks such as SARS or the H5N1 (Bird Flu), it appears to be significantly more contagious and harder to control.
Other Recent Data
Two recent studies are important to highlight . The first, The role of absolute humidity on transmission rates of the COVID-19 outbreak suggests that it’s unlikely that warmer summer temperatures will slow the spread of the virus as some initially hoped. And the second, The Novel Coronavirus, 2019-nCoV, is Highly Contagious and More Infectious Than Initially Estimated is the more recent study estimating the R0 of the virus suggests that the containment measures in China have only reduced the R0 to somewhere around 2.3–3.0. Indicating that the virus is still spreading rapidly within China, and the drastic containment measures will eventually need to be taken by other countries as well.
In my initial analysis of the COVID-19 outbreak (then Coronavirus) I outlined 4 different scenarios for the outbreak. See the full writeup here. After some thought, I’ve expanded the possible scenarios to six cases:
Case 1 — The Optimistic Scenario
In the optimistic scenario, the CFR of the virus will be perhaps .5% or below. Moreover, the spread of the virus outside mainland China will stabilize and we will not see additional community spread in other countries. This would be a good indication that China’s containment measures worked and the R0 value has fallen to 1 or bellow. In this case China will probably make a quick economic recovery.
Case 2 — Deadly but Containable
In this scenario the virus would have a high CFR, say above 3%, but containment measures are able to slow the spread of the virus and the R0 drops to 1 or below. In this case outbreak would be closer to the 2003 SARS outbreak.
Case 3 — Pandemic with Low CFR
In the case three scenario we will be able to calculate that the CFR of the virus is fairly low. Somewhere between .1 % and 1%. However, the virus is not contained to China and we see sustained community spread in multiple countries outside of China.This means that the virus will likely spread around the world — but it would be about as bad as the flu or perhaps 10X as bad as the flu. At the low range of this scenario the economy would make a fairly quick recovery and at the high range it could be a moderate drag on global growth. This could be one par with the 2009 Swine Flu pandemic or slightly worse.
Case 4 — Pandemic with Moderate CFR
In this scenario the virus is not contained within China — and the CFR is somewhere around 2% maybe a few percentage points higher. Anecdotally this appears to be the current base case assumption among the experts. This would almost certainly have a fairly large economic impact. Hospitals might be overwhelmed, people would be wearing masks on the street. Millions could die in the United States and other developed countries. Look for several countries to be pushed into recession.
Case 5 — Pandemic with High CFR.
In the case five scenario, the virus is not contained in China, and it ends up having a CFR somewhere between 4.5%–8%. This wouldn’t be quite as bad as the 1918 Spanish Flu, but would probably have an enormous economic impact — particularly given the globalization of the economy. Look for draconian quarantine measures to be put into place around the world.
Case 6 — The Big One
In the worst case scenario the virus is not contained in China, and it ends up having a CFR or somewhere above 8%. Maybe the fatality rate is slightly worse than we hope — and healthcare systems become overwhelmed causing the mortality rate to rise substantially. In that case, this is The Big One.
Based on the current estimations, we appear to be on track for scenario four. Obviously more data and containment measures could show that the virus and not that severe, although arguably overwhelmed healthcare systems could easily push the world into Case five.
For daily updates, check out the coronavirus scenario tracker.
Or on twitter: @johsuafkon