We are so early into the outbreak of COVID-19 it is dangerous to jump to many conclusions based on the very limited data we have. We have studied influenza for decades with a universe of hundreds of millions of cases. Yet, there is still a great deal of uncertainty regarding influenza’s fatality and transmission rates. With COVID-19, we are dealing with a data on only about 150,000 cases over less than three months. On top of that, we know little about how accurate the data we have is.
But, with that huge caveat in mind, there may be a pattern developing based on the experiences of China, South Korea, Italy and Iran. These are the current hot spots. All of the other countries that have reported cases have fewer than half the cases of these four countries and their outbreaks started somewhat later.
The World Health Organization (WHO) began publishing daily updates on the number of cases by country on January 21.1 According to their data China already had 278 cases at that time. There were only four cases outside of China. The outbreak in China rapidly accelerated and by the first week in February, it was reporting over 3,000 new cases per day. The number of cases in South Korea, Italy and Iran began to accelerate about two weeks later.
The number of new cases in China and S. Korea peaked after about two weeks and then began to decline almost as fast. Yesterday (March 15) China and S. Korea reported only 27 and 76 new cases, respectively.
But Italy’s and Iran’s experience has been different; Italy’s starkly so. In both countries, the number of new cases has continued to climb, hitting new highs yesterday of 3,497 and 1,365, respectively. And, discouragingly, both are well past the two weeks that it took cases in China and S. Korea to peak.
Once again, we should be very hesitant about drawing too many conclusions based on the very, very limited data we have. There is undoubtedly wide variance in accuracy and methodologies of how the data on new cases is being collected. There has also been a fair amount of justifiable skepticism about China’s reported numbers given its penchant for secrecy and gilding the lily about anything having to do with its country.
Nonetheless, there may be some inferences to be drawn.
First, the increases in the incidences of infection has been not “exponential” anywhere, notwithstanding the media’s repeated characterization to that effect. I frequently see charts showing the cumulative numbers of cases. Those are always going to show a much steeper increase and resemble an exponential function. But the critical parameter is how fast the spread of virus is accelerating or deaccelerating. That will be reflected by the number of new cases over time, not the number of cumulative cases.
Second, it appears that China and S. Korea were able to contain the virus in about two weeks. As nearly as I can determine from media reports, each had somewhat different containment strategies, but both took extraordinary measures. But Italy, and Iran to a lesser extent, have not been able to contain the outbreak in their countries. Italy’s numbers are particularly alarming given they have now locked down the entire country and the virus is still spreading at a rapid (but not exponential) rate. It would be very helpful to determine if China and S. Korea did something(s) that Italy and Iran have not.
Third, the overall infection rates have been extraordinarily low. In Hubei province where the outbreak started, only 1 in 870 people have been infected so far and the number of new cases has slowed to a trickle. In other words, even at the epicenter of the outbreak where it is has now largely been contained (at least, according to the Chinese), only slightly more than one/tenth of one percent of was infected. 2 The rates have even been lower in the other countries (S. Korea=1:6,300 – Italy=1;2,900 – Iran=1:6,600).
I offer this perspective on the data not to suggest that anyone is overreacting or underreacting to the outbreak. But, when facing a new threat, we have been programmed by evolution to exaggerate danger in our minds, at least until we can accurately assess the degree and lethality of the new threat. That instinct can only be overcome by dispassionately examining actual data.
This is a link to the spreadsheet I have been working on. You are welcome to download it and do your own analysis. If anyone finds any errors in it, please let me know. I will be regularly updating this analysis.
1 All of these statistics come from the World Health Organization’s situation reports. On February 17, the WHO changed the methodology for counting new cases from China’s Hubei province. Prior to February 17, the WHO had only included laboratory confirmed cases, not clinically diagnosed cases. On February 17, WHO added the previously clinically diagnosed cases (≈ 17,688 cases) in a lump sum. For charting purposes, I have “spread” the clinically diagnosed cases over each previous day proportionate to the laboratory confirmed cases from China for that day.
2 The actual number of infections in Hubei was undoubted higher than .1% as asymptomatic and mild cases likely went undiagnosed. But even if the infection rate was ten times higher than reported, that sill means that 99% of the population was never infected.