Coronavirus: numbers and hindsight

Note on 10 April 2020: I first wrote documented my thoughts on COVID-19 on 25 February 2020. This article has been updated to reflect how the situation has changed over the past 6 weeks.

I first started seriously worrying about COVID-19 in mid February. At the time, Australia had 15 confirmed cases (all from overseas) and the infection rate was 0. That is, there were no new cases developing each day.

However, what peaked my interest and spurred my nervousness was the number of countries that were reporting their first reported case. At the time, I was using Rami Krispin’s fantastic R package to pull the latest COVID-19 figures from Johns Hopkins University (their geospatial dashboard is linked here). This was one of the first sources available to the public.

On 25 February I wrote:

One of the difficulties in understanding the impact of coronavirus in different countries is just how quickly the number of cases keeps changing. Each day there seems to be more countries at risk, with a higher number of cases.

In hindsight, this statement was mainly because the charts I was making started to have a longer and longer x axis.

There was a long tail of countries with single digit cases, however it seemed to grow longer each day. I remember seeing Afghanistan and Papua New Guinea pop up on the list. This is when I knew global health infrastructure was going to be tested.

When I pulled together a chart of the most recent confirmed cases, I wrote:

China has by far the largest number of cases worldwide. There was a significant jump in confirmed cases after the diagnosis method was changed in mid February. The growth in the number of new cases has slowed in the past 10 days, and the number of recovered cases is growing.

South Korea has become a major infection area (however the nation still only has a fraction of the confirmed cases in China).

The main risk here is that growth in South Korea and Italy has been rapid in the past 7-days, suggesting it has been multiplying in an uncontrolled fashion. This also aligns with research of high R0 values (e.g. each person infecting 2–3 others) in South Korea.

The power of hindsight is powerful here, and I have a trove of charts I made about the risk of Thailand and Hong Kong which now seem silly compared to the carnage occuring in the US and Spain. My naivety of what was to come is emptimised in my summation:

Just 10 days ago (16 February) the coronavirus story was very different. Other Asian countries (including Australia) looked to be the most at risk, and since then we’ve had 12 new countries report their first confirmed cases. Australian confirmed cases remains low, and recovery is high. While Australia had some of the earliest cases outside mainland China, our number of cases has remained low since the travel ban was implemented in January.

When I first started investigating and documenting COVID-19, every country outside China has less than 80 confirmed cases. Perhaps because of this, my summation of what was to come seems jubilantly optimistic given the current global situation in mid-April.

Whether this means a widespread pandemic is imminent, or a prolonged period of uncertainty is more likely it may be too early to tell. However a data oriented view helps to put the virus in perspective, and obtain a day-by-day perspective as the crisis unfolds.

For the first time in my adult life, I was truly standing at the edge of history, and managed to glimpse a peek at the other side. Perhaps I still am at that same edge, but I stand by my initial observation: it may be too early to tell. History, after all, is a product of hindsight.