Success & Failure — Data from 7 Countries

Periodically, I think it’s a good idea to compare the U.S. overall to other countries around the world. I’ve been following 7 countries since the pandemic began, and an update is due.

The metrics readily available and uniform are very limited. There’s no testing data, no hospitalization info, no Rt calculations, etc. We have cases and fatalities — that’s it. With that in mind, I think the following metrics are important. I’ve also indicated the importance of the metric for me personally as I assess it, on 1 to 5 scale, with 1 being the most important. I’ve also included the ID of the chart that shows the metric for all countries I’m tracking.

  • Prevalence Index [E1]: this is the population divided by the number of cases. It’s a normalized indicator that allows you to visualize how many people you would have to gather together in a locality to find a single direct encounter with CV19 — a positive result on a test, an illness, a recovery, or a death. In other words, a case. The prevalence index also gives a very general indication of how far a locality is from herd immunity. Most experts say that herd immunity for CV19 would mean at least 70% of the population would have to have antibodies to CV19 to achieve herd immunity. That would mean the prevalence index would be around 2.

Metric importance: 2.

  • Cumulative Cases per 1M People [E2]: another normalized statistic that allows you to compare outbreaks in localities with very different populations. This is a good general gauge of how well a country has done in managing the outbreak. It’s a good proxy for the prevalence index, since a country that’s very low on the prevalence index will generally have a very high cumulative cases per 1M.

Importance of this metric: 4.

  • Daily New Cases per 1M [E3]: This is a good indicator of whether the outbreak is growing or retreating. It’s normalized so that one can compare localities with different populations. It’s also based on a 7 day moving average to filter out some of the noise from daily data. Importance of this metric: 3
  • Case Doubling Days [E4]: Without an Rt (current rate of transmission) value, this is the next best thing. In the initial stages of an outbreak, doubling days are often in the single digits. As case and fatality counts rise, we should see the doubling days go up as well, even if cases are going up a lot. When the doubling day plateaus or worse, declines, that spells trouble in most cases. It’s an indication that the outbreak is in danger of getting out of control.

Importance of this metric: 2

  • Cumulative Fatalities per 1M People [E5]: See Cumulative Cases per 1M above [E2]. I would only add here that there is a bit less “fuzziness” in the data with fatalities. Cases are based on testing and direct observation, and in some countries, testing is severely limited. In Germany, for example, the case count might understate the number of infected people by 10%, but in Brazil, the number is more likely to be something like 50%. Fatalities are under-reported as well, but not to the extent of cases.

Importance of this metric: 3

  • Daily Fatalities, 7 Day Avg. [E6]: Same rationale as Daily New Cases above [E3]. The same differences between case data and fatality data apply. The difference here is that this is not a normalized number, so it gives you an idea of the scale of the pandemic in absolute numbers. A country where 1,000 people a day are dying is in a very different state, even if it’s a very large country, than one where 25 people a day are dying.

Importance of this metric: 2

  • Daily Fatalities per 1M People [E7]: Same rationale as Daily New Cases above [E3]. The same differences between case data and fatality data apply.

Importance of this metric: 1

  • Case Fatality Rate [E8]: This number is a percentage: fatalities today, divided by cases 7 days ago. It’s an imperfect indicator for many reasons. First and foremost, a country like Mexico, with very low testing rates, will tend to have a much higher CFR. Second, the case count includes unique individuals, total tests, and serological tests, at least in the U.S., so that drives down the percentage. But over time, we should expect that number to fall in each country as testing increases and also as doctors get better at saving people.

Metric importance: 3.

  • Fatality Doubling Days [E9]: See Case Doubling Days [E4] above. Again, the difference between cases and fatalities in “fuzziness” is important here. A number less than 30 days indicates serious problems, not just in containing the pandemic, but in controlling the number of people dying from it.

Metric importance: 1.

When I develop a Zorgi Score for the country level, I’ll probably use these metrics and weight them by the importance assigned to each. For today’s update, I’ll briefly cover each country and assign a “gut” Zorgi Score. I just want to emphasize this is not a public health statistic, or a travel advisory score, or an epidemiological projection. It is simply an indicator of how serious I personally think the situation is in that country based on the metrics outlined above. If you don’t like the Zorgi Score, great. Ignore it, and just look at the data.

Denmark — Zorgi Score: 0.5.

Denmark has a prevalence index of around 500. That’s lower than Mexico, but the difference is that in Denmark there was a lot of testing very early on, while in Mexico there was hardly any. E2 shows an almost perfectly flat line from the middle of May to today, only copied by Germany. Daily new cases are miniscule, around 5, and there have been several days with no new cases at all. Case doubling days are at 74, only exceeded by Germany at 83. Denmark and Germany have the lowest cumulative normalized fatalities as well. You can see how well Denmark has contained the pandemic on E6, where the line goes down to 0. This is also dropping the CFR [E8] down to 4.8%. Finally, the fatality doubling days [E9] is at 76, the highest of all 7.

Germany — Zorgi Score: 1.0

Prevalence index almost identical to Denmark. Cumulative cases per 1M [E2] shows an upward trend, unlike Denmark, because there are still cases popping up here and there. But the logarithmic scale hides the immense difference between the rise in Germany vs. the rise in the US or Brazil. In every other respect, Germany follows Denmark.

Sweden — Zorgi Score: 8.0

Sweden’s prevalence index [E1] is just above that of the US, at 149. In normalized cases [E2], it’s one of the highest at around 6,100 — right up there with the US and Brazil. Normalized daily new cases [E3] took a huge jump in the last two weeks of June, but then dropped down to 106, still one of the three highest. A very serious indicator is doubling days [E4], where it flattened out and even declined over the last 2 weeks. Sweden was held up as a “model” for lockdown protesters, until people got a good look at their normalized fatality rate [E5]. Here it’s one of the worst in the world, at 521, compared to a still miserable 391 in the U.S. Daily fatalities are fortunately dropping now, going from around 60 two weeks ago to 25. The CFR [E8] is much higher than we’d expect from a modern European country, simply because the death toll per 1M is so high. The good news is that fatality doubling days [E9] are rising steadily.

Brazil — Zorgi Score: 22

There’s nothing very encouraging about Brazil’s numbers. Prevalence is at 123, a bit above the US, but dropping rapidly. The rate of case growth [E2] is much steeper than any other country, and may surpass that of the US within a couple of months. This is underlined by the normalized daily new cases [E3], where Brazil is #1, with 175, compared to 1127 in the U.S. Case doubling days are at 21, the lowest of all the countries. Cumulative fatalities per 1M [E5] started off low in May, but now surpass Ecuador and are challenging the U.S. Daily fatalities [E6] are one of the highest in the world, at 984. The CFR [E8] is low, but both cases and fatalities are, by most reports, severely under-reported. Finally, Brazil has the second lowest fatality doubling days [E9] of all the countries.

Ecuador — Zorgi Score: 13

Ecuador started off with a very serious outbreak, but locked everything down. Their government, unlike Brazil, did not try to deny the seriousness of the pandemic, at least once they realized it was growing out of control. Normalized cases per 1M [E2] are rising, but not at nearly the same level as Brazil, Sweden, and the U.S. Daily new cases per 1M [E3] have leveled off to around 45. Case doubling days are out of the danger zone at 54. Daily fatalities per 1M [E7] are way below that of Mexico and Brazil, at 2.4. Finally fatality doubling days [E9] is out of the danger zone at 49 days.

Mexico — Zorgi Score — 18

Mexico started off poorly by virtually ignoring the pandemic. The prevalence index [E1] is still much higher than other countries, but that’s primarily because of the lack of testing. Every metric involving cases in Mexico needs to be second-guessed because of this. That’s why the case doubling day number [E4] of 21 is so serious. Cumulative fatalities per 1M [E5] are rising on a trend line looking like Brazil’s. Daily fatalities [E6] at 689, are surpassed only by Brazil and the U.S. On a normalized basis [E7], daily fatalities are the highest of all, at 5.5 per 1M people. This results in the highest CFR [E8] of all the countries, at 13.9%. Finally, fatality doubling days are at a very low 21 days, the lowest of all the countries.

U.S.A. — Zorgi Score — 18

The US, like Brazil, tried to wish the pandemic out of existence during February and March, and locked down only when it was clear that things were out of control. As a result, the US has the lowest prevalence index of all the countries, at 123. Basically this means that a statistical extended family in the U.S. has direct experience with 1 case of CV19. Our number should be around 500, but this is where incompetence gets you. In normalized cumulative cases [E2] we’re the leader of the world, at 8,114. Our normalized daily new case number [E3] leveled off during the lockdown, but over the past two weeks has gone up at a sharper rate than any other country. Case doubling days [E4] for now are high, but that’s largely a function of the huge number of cases. In cumulative fatalities per 1M [E5] we are second, at 391, only surpassed by Sweden at 521. Daily fatalities [E6] dropped for 10 days or so, but are now back up to 930, just a hair below Brazil at 984. Fortunately, the fatality doublind days number is around 60, but that may level off if the daily fatality rate doesn’t drop soon.

Thank you all for your great comments, suggestions, and error corrections. I even appreciate comments from people who totally disagree with me, as long as they’re not just trolling. I especially appreciate offers to buy me a cup of coffee, a burrito, etc. But instead of giving something to me, I’d like you to consider giving to the Equal Justice Initiative.

You can donate here.

Stay healthy and safe, everyone!

Leave a Reply

Your email address will not be published. Required fields are marked *

 

This site uses Akismet to reduce spam. Learn how your comment data is processed.