Hoping That Numbers DO Lie

A few days ago, a colleague who has proven expertise in statistical analysis texted me:

This data from China is so obviously made up that it insults my intelligence. They could have at least spent a little time trying to make it believable…

Of course, I hoped he was employing hyperbole. After a closer look at the numbers, I understood his skepticism.

Note: If you are interested in What the COVID-19 Data Is and (Is Not) Telling Us, look for a new webinar by Ralph Smith later this week.

Countries throughout the world are struggling to contain the virus. Those that were impacted first (China, Iran, Italy) were at the most disadvantage. A deadly and highly contagious disease was spreading for weeks before they could take defensive measures.

Due to its size and large, distributed population centers, China was by far at the most risk. Even if they could contain the virus to Wuhan (11.9 million people) or Hubei Province (58.5 million), the challenge of reducing the spread would have been daunting. But the virus spread to at least 25 other provinces.

And yet, key metrics (new cases, hospitalizations, deaths) have improved by an order of magnitude from one month to the next. Compare the growth of cases in China to Iran and Italy:

Clearly, the scales are different but the “curves” are still instructive. China had a single primary peak and then dropped to less than 100 new cases for 15 days in a row. Any mini-spikes it attributed to international travelers. This is an unusual curve.

Iran and Italy show more typical curves with multiple peaks and plateaus over time. Italy is stepping down from levels of 6,000 to 5,000 to 4,000.

Making the numbers China is reporting more dubious is the R0 value or “contagiousness” of COVID-19. An average of multiple studies pegs the COVID-19 R0 value at 3. In other words, each infected person will infect three other people. We saw how devastating that can be with the first document case in New York City. One sick passenger from Italy quickly became hundreds of sick people in the suburb of New Rochelle.

Absent a vaccine, it typically takes several months to bring the effective R0 rate below 1.0. It requires both mitigation (social distancing, etc.) and the development of “herd immunity.” That is simply not possible in a matter of weeks.

So, either:
a) China has done a miraculous job of containing COVID-19 via conventional methods in the world’s most populous country, or
b) The data from China are wrong, or
c) The Chinese have some unknown methods of treating the virus.

I’d like to believe the answer is a or c but prudence drives me toward “b” as in Bad Stats. I recommend extra caution if you have China travel in your 2020 plans or if your supply chain depends upon production in mainland China. (It is worth noting that Taiwan has done a miraculous job containing COVID-19.)

Stay safe and healthy.

–Paul King

Pinocchio is copyright by The Walt Disney Company.