This is interesting. The provocative part is not the data, nor as most of the rep;ones in this thread focus on, estimates of CFR but the striking difference between US and Europe

Important to spend some time thinking about the origin of the difference https://twitter.com/vprasadmdmph/status/1248056240581693440

The proportion of deaths that are people <65 is almost 10x higher in NYC than European epicenters. But this raw figure doesn’t correct for the age distribution of the whole population - New Yorkers likely much younger on average than citizens of Lombardy

So younger people are bound to be a higher percentage of those infected and hence if those dying. It’s also possible poorer younger people live at highest density in US whereas in Europe it is the elderly who live at higher density (eg in care homes)

BUT these factors unlikely to account for the magnitude of the difference - 2x greater proportion of younger people among NYC deaths, maybe. But 8x? Improbable. So what else may be going on?

Differences in healthcare provision

Younger people’s outcomes may be much more subject to treatment - so if treatment is more effective on Europe than US it could account for lower proportion of deaths in younger people. If that’s true, it deserves urgent consideration since it is a modifiable factor

Unfortunately both under- and over-treatment could be accounting for the difference.

Without comprehensive healthcare coverage, how many poorer, younger New Yorkers are not getting access to critical care beds when they need them?

Alternatively, US docs have a well-documented tendency to over-treat. We are learning that early ventilation and particularly PEEP, may be very detrimental causing lung damage on top of the virus - when patients ability to breathe is largely unaffected

#COVID19 lung disease is oxygenation failure rather than respiratory failure, perhaps due to microvascular thrombi but in any case with a large ventilation:perfusion mismatch. Are US docs intervening more aggressively than European counterparts with detrimental consequences?

Maybe this learning evolves with experience - NYC critical care docs coming to the same conclusion - so maybe European data is just more “mature” - young people die disproportionately EARLY in the epidemic, while local docs learn how best to keep them alive

If so, it becomes a matter of urgency to ensure best practice is disseminately widely, quickly rather than just rushing to increase critical care and ventilator capacity

Addendum: also wonder if national systems in Europe are quicker to spread best practice in rapidly-evolving new circumstances than the fragmented private provision in US