Russia has been running at ~100 deaths per day for the past three weeks. As most everywhere (~60% across 14 analyzed countries), this is an underestimate – in Moscow, by a factor of ~3x.
Something that we can tell when the gross mortality stats became available a week ago, suggesting ~2,000 excess deaths vs. 642 official deaths from COVID-19.
“Not great, not terrible.” Much worse than East Asia. Worse than Visegrad. But still better, per capita, than the US and most of Western Europe – at least to date.
However, if there’s one region where the data is off by an order of magnitude – as has been intermittently alleged – it is the North Caucasus.
A couple of days ago, Dagestan’s Health Minister said that 700 people had died of community acquired pneumonia in Dagestan, including more than 40 doctors. The official figure is 32 deaths (as of today).
In an open source project to track doctor deaths from coronavirus, Dagestan is far in the lead, with 38/241 names (as of today and the time of writing). For comparison, Moscow, with an official population that is four times greater and, with 1,651 deaths as of today, commonly recognized as the epicenter, accounts for 71 of the names. In per capita terms, Dagestan’s epidemic may even be worse than Moscow’s. (Or perhaps their hospitals have an exceptional lack of PPE, with their doctors getting infected at much higher rates).
But more likely there is some extreme undercounting going on it Dagestan.
Other badly affected regions:
- Moscow – 70/241
- Moscow oblast – 44/241
- Dagestan – 37/241
- Saint-Petersburg – 20/241
- Chechnya – 4/241
- Krasnodar krai – 3/241
- Leningrad oblast – 3/241
- North Ossetia – 3/241
- Nizhny Novgorod oblast – 2/241
- Ingushetia – 1/241
- Other regions – 54/241
Although Dagestan is a clear outlier even there, the other two DICh republics are also overrepresented, with 4 doctor deaths in 1.4M Chechnya, 1 in 0.4M Ingushetia. Also hard hit is North Ossetia, with 3 doctor deaths out of 0.7M.
This is much higher per capita than in the ex-megapolis Russian heartlands, e.g. 3 deaths in 1.7M Leningrad oblast (which surrounds Saint-Petersburg), 2 deaths in 3.3M Nizhny Novgorod oblast, and 3 deaths in 5.2M Krasnodar krai (which is a popular tourist destination, hosting Sochi).
“Other regions” has about 110M people, which generated 54 doctor deaths. So, about 2 per million.
The Business Insider article that reported this story in English suggests that if all Russia was undercounted like Dagestan there’d be 50,000 deaths, which already makes Russia as bad as the US.
If the overall rate of infection and death in Russia followed the same pattern as Dagestan’s, Russia could have actually been hit by more than a million cases and already suffered more than 50,000 deaths from the pandemic.
But this would also suggest that all Russia is as bad as Moscow (2,000 real deaths*10 = 20,000, and perhaps 2.5x as much in the nineteen days since), where we know the situation quite well thanks to the April figures for gross mortality.
This is not very plausible given its status as epicenter, as well as the doctor deaths stats above.
So how many actual deaths in Russia? Officially, Moscow accounts for 1,651 / 2,837 as of today. If we assume a 2-3x undercount in Moscow based on the discrepancy between official COVID-19 deaths and excess mortality in April, the real figure in the capital is 4,000 by now. If in turn we then assume that doctor deaths are on average equally likely to be reported to that doctor deaths database across regions, where Moscow accounts for ~30% of deaths – twice lower than 60% in the official figures (1,651/2,837) – then we may guesstimate that Russia’s total numbers of deaths right now is perhaps 13,000, or a fourfold discrepancy with official figures.
This is of course very back of the envelope stuff. But it’s clear that deaths are substantially higher than official figures, while being lower than wildest extrapolations.