Coronavirus: Not Even Italy Is Like Italy

As the coronavirus crisis unfolds, will the US be the next Italy?

At last Tuesday’s White House press briefing (31 March), the government’s top medical experts Dr. Birx and Dr. Fauci outlined the model forecasts behind the decision to extend the economic shutdown: the coronavirus could cause some 2 million US fatalities if nothing is done, and at best 100,000 in a full mitigation scenario.

The model projections are informed by Italy’s experience, said Dr. Birx. Two days later she added: you’ve seen the US curve and Italy’s curve, we have to bend our curve, we don’t want to end up like Italy.

Every time I hear this, my reaction is “but…not even Italy is like Italy!” But more on this shortly.

Dr. Fauci stressed he does not accept 100,000 fatalities as inevitable, and strongly hopes the actual number will be considerably lower. One reporter asked: Dr. Fauci, I don’t understand, if 100,000 is the lowest projection for a full mitigation scenario, doing everything we possibly can, how do you think we can get something lower? Dr Fauci’s answer ran something like: this is a very good question, and yes, the model projects at least 100,000 in a full mitigation scenario, but you know, models are very sensitive to assumptions and this is where we have to ask, is the model really giving us a good indication?

I discussed the pitfalls of these models in my previous blog — and as the economic shutdown starts destroying the livelihoods of millions of people, we should keep them in mind.

But now let’s talk about Italy.

The covid outbreak in northern Italy has had a massive psychological impact across the western world, and governments in the US, UK, France and elsewhere have followed Italy’s lead in shutting down their economies.

But country-wide average statistics for Italy are always misleading, whether you are looking at income levels, productivity, unemployment — or, as it turns out, covid-19. This holds some important lessons for how to read and counter the current health crisis.

Lombardy has been hit far worse than the rest of the country. As of April 3rd, Italy has recorded nearly 120,000 Covid-19 cases and close to 15,000 deaths. Lombardy accounted for 40% of cases and nearly 60% of deaths; a disproportionate share given that the region is home to less than 17% of the country’s population.

To compare Italy to other countries, it is useful to see what happens if we separate out Lombardy. The next chart shows the increase over time in the number of confirmed cases per 1,000 inhabitants — to adjust for different population sizes. Lombardy is way at the top, with 4.7 cases every 1,000 people. Italy ex-Lombardy (green dots) has less than one third of that number, at 1.4 — fewer than Spain and not too far from France and Germany or the US.

Excluding Lombardy, Italy’s mortality rate is a still high 8.8%, on a par with Spain — though now lower than France. But the shocking fact is that if you contract covid-19 you are twice as likely to die in Lombardy as in the rest of the country: The mortality rate in Lombardy has climbed to a horrifying 17% — nearly one in five confirmed infected persons has died.

To summarize: Taking Italy as a benchmark can be misleading, since its data are so heavily skewed by Lombardy. Once you exclude Lombardy, Italy shows an incidence of infection broadly in line with other large European countries and a mortality rate at the high end of current European estimates, but well below the Italy-wide average.

Moreover, in my view the evidence does not suggest that the US is following in Italy’s tracks.

Those who say the US might become like Italy often focus on the absolute number of confirmed cases. The US now has over twice as many confirmed covid-19 cases as Italy; it has also conducted more than twice the number of tests. And remember that the US population is over 5 times larger than Italy’s. To compare the two countries we need to adjust for population size. Italy has tested a larger share of its population than the US so far. But when Italy was at the stage of testing where the US is now, it had a higher share of infected people and a mortality rate four times higher than the US today — the chart below suggests the US is on a very different mortality curve than Italy’s.

Within the US, New York raises the greatest concern, with its health care system already under strain. New York State’s positivity rate (confirmed cases / tests) is about 38%, on a par with Lombardy’s peak. On the other hand, the mortality rate in New York is still under 3% after about one month of testing; Lombardy’s rate had jumped to 8% already by early March.

The US is not like Italy; New York is not like Italy. Not even Italy is like Italy.

And encouragingly, the latest figures seem to suggest that even Italy might be turning the corner. Fingers crossed.