Analysis of covid-19 pandemic per country using deaths per million

This is analysis on data so far, not a prediction model.

Update 7.4.2020

First graph: estimate each country’s lag vs. Italy, by shifting the curve so that <deaths per million> is around 6. 

Second graph: same as above, but for full scale of y-axis. Straight comparison of each pattern. Country order in line with severity

 

The objective of the analysis is to compare the expansion of the covid-19 disease in various countries in the pandemic caused in early 2020, emanating from China.

How is this analysis different to others.

Two choices are required.

First, I select the <ratio of number of deaths over a million population>, so that the comparison keeps as many factors as possible to be the same among different countries.

Second. In order to align the different development stages of the pandemic in each country, I avoid using: the day of the first confirmed case; the day of the first death; the day of the third death; or the day of the tenth death. All these, I believe are subject to either random phenomena because they may come too soon, or a result of different methodologies applied, e.g. number of tests. I align the curve of a country to Italy, on the day that the variable above <deaths per million> had the same value. This value is not too small. For most comparisons is around 5-6 deaths per million. So after that, the comparison is made on how different was the expansion of the pandemic in a country compared to Italy.

Hence, I first estimate the lag, in number of days, that a country has the variable equal to around 5, with Italy. For example, if the lag is 10 days, it means that this country reached 5 deaths per million 10 days later than Italy. The graphs show the development from that stage. If development is greater than Italy’s (for example is Spain), the situation is worse.

The same graph is presented twice. The first graph confirms the estimation of the lag, so the scale is typically between 0 and 10 (deaths per million), to justify that around 5 the two curves have the same value. The second graph shows the development of the curve.

 

 

Two additional charts. No lags to any country. First with deaths per million and second confirmed cases per million.

 

Results:

Lags (approximate: +- 1)
Spain: 7 days
Belgium: 13 days
Netherlands: 12 days
France: 12 days
UK: 16 days
US: 20 days
Sweden: 17 days
Luxembourg: 11 days
Switzerland: 12 days
Portugal: 18 days lag
Germany: 20 days
Iran: 4 days
Greece: 20 days

Conclusions (as of 3.4.2020)

1. Spain is worse than Italy
2. Belgium is worse than Italy, and better than Spain. I haven’t seen any focus on Belgium, because the absolute number of deaths is relatively small. However, the ratio of deaths / million is growing at a faster pace than in Italy at the time. So, the stress to the national system is greater.
3. UK, France, Netherlands, and the US seem to follow Italy’s curve. The data for the last three days for the UK and France show a deterioration of a previously slightly milder picture than Italy. This is a key point to see how this trend will continue. It could be a little upgear, which will come off and the curve will continue lower than Italy’s, or it will remain and things may become slightly worse.
4. Sweden, Luxembourg and Switzerland show a slightly better pattern than Italy.
4. Germany follows Iran, in a clearly better condition than Italy.
5. Greece shows a much better pattern than all the above. The lag of Greece was estimated when the value of the variable was around 2-3 deaths per million, as only its current value is around 5.

Good news:

Italy’s curve seems to be stabilised. In the last 10 days, the daily number of deaths fluctuates around 760 with no tendency to be increased. This can be the inflection point. The lags per country above, may indicate distance from entering the second phase of stabilisation. It is hard to predict how long this stabilisation phase (or peak) will last. Don’t assume it will be a couple of days. After that, the third phase will be to have reduced daily deaths.
Spain also shows signs of stabilisation, but only during the last very few days. We need to wait and see how things will develop.

All data are up to 3rd April 2020, unless stated differently.

 

Update: 17.4.2020. More in depth analysis, daily graphs in the new website with this topic http://covid.antibaro.gr 

 

Andreas Stalidis

antibaro.gr/article/26507