With the electoral campaign in Madrid, the comparisons of the covid figures by regions multiply. They are used to draw firm conclusions about the success of certain measures or certain managers. But it is a complicated exercise even using exact and true data: with a phenomenon like the pandemic, depending on which you look, you can project a different image. “It is a complex problem in every sense of the word, with a multitude of factors that are related to each other”, sums up Saúl Ares, a researcher at the CSIC. It is the same thing that Clara Prats, from the Polytechnic University of Catalonia, underlines: “There are many variables that intervene and the way in which they are related.” And the epidemiologist Alberto García-Basteiro also emphasizes this: “There are many factors and their interrelation is complex: only very sophisticated models could try to estimate the fraction attributable to each determinant.”

What we can show are the numbers of cases, income and deaths. We start with the death statistics, which is the most significant measure of the impact of the pandemic, showing those registered since March 2020, in each wave and per thousand inhabitants of each province.

The table collects the total number of deaths in each of the 50 Spanish provinces (in addition to Ceuta and Melilla) since the beginning of the pandemic until mid-April of this year. For the second and third waves we use the confirmed death figures collected by the Carlos III Health Institute. For the first, on the other hand, we used the excess of deaths collected from the INE, because the detection of coronavirus cases was very bad then.

More information

According to these data, the provinces with the most deaths from covid since the beginning of the pandemic are Soria (5.9 per thousand inhabitants) and Segovia (5.6); followed by Cuenca, Ciudad Real and Salamanca (around 4). Then Avila, Guadalajara, Toledo, Madrid, Albacete and Palencia appear (with more than 3 deaths per thousand inhabitants). At the opposite extreme appear the island provinces, Galician and some Andalusian, where deaths are a tenth that in Soria or Segovia.

But the table also shows differences by waves. This is the case of Madrid, for example, which appears as one of the worst in total (the ninth) and in the first wave (the seventh), but also in the middle of the second wave and among those with the least deaths in the third wave (the fourteen with less).

As we will see, these numbers vary when we begin to consider factors such as the age of the people or the number of people susceptible to becoming infected.

How much does the age of the population matter?

Covid-19 is a disease that prevails in the very old (almost 90% of those who died since March 2020 were over 70 years old). Therefore, the provinces where there are more elderly faced a greater danger. To take this into account when comparing territories, the mortality figures for each province can be adjusted according to their age structure: calculate what that mortality would be if there were the same proportion of young and old as in all of Spain.

This table collects the adjusted mortalities for the total number of deaths in the pandemic and for those who have died since December 2020, what we call the third wave.

This setting does not alter the overall image, but it does introduce some nuance. The most important thing is that it improves the statistics of provinces with a much older population, such as Soria, Segovia and Ciudad Real: a part of their high mortality was due to their aging population, although they remain three of the four regions with the worst figures also after to adjust.

Other provinces such as Ávila, Salamanca or Cuenca also see their figures decrease. In fact, Madrid now surpasses them and becomes the fifth province with the worst mortality. Guadalajara – it goes from seventh to second -, Barcelona – goes from being in the average to being the ninth worst – or the Valencian provinces, which have younger populations, also worsen. Almería or Cádiz, which were among the territories with the lowest mortality, now come out in half.

If we look at the third wave, Madrid is still among the provinces below the average in mortality now that we have corrected the advantage it has for the age of its population.

How much does the impact of the first wave matter?

The epidemic tends to slow itself down. It does so through two mechanisms: on the one hand, because each infected reduces the number of people susceptible to reinfection. And to that is added a possible harvest effect: When it appears, the virus is more lethal because it finds more weak or vulnerable people. This means that, hypothetically level playing field, deaths will be lower in the worst hit areas.

But the available data limit the importance of this phenomenon in Spain. Having suffered a very hard first wave has not prevented some regions from also having many deaths in the second and third waves. Nor has avoiding the first necessarily meant a worse second or third wave.

As can be seen in the graph, there are provinces such as Lugo, Ourense or the Canary Islands that have had few deaths both in the first wave and in the successive ones. But even ignoring those, which are probably protected in part by their geographical position, the relationship between first wave and subsequent waves is weak. There are some provinces that were among the worst in the first wave and then no longer, such as Madrid, Ciudad Real or Segovia; others avoided the spring outbreaks and then suffered in winter, such as Palencia, Zaragoza, Teruel or even Almería; but there are also some provinces with always high mortality in all waves, such as Soria, Guadalajara and Toledo.

Another way to judge the mortality figures in the second and third waves is to adjust them for the number of people who were susceptible to being ill at any given time. This adjustment can be made using the seroprevalence study of the Ministry of Health. For example, if it is estimated that in a province 20% of the people passed the virus before December, there will only be 80% susceptible, so that mortality is higher if we calculate it on them and no longer on the entire population . The following table shows the mortalities by province with this adjustment, and also the previous age structure adjustment.

This new adjustment introduces new nuances, but no essential change in the global image of the pandemic. The provinces with the most deaths in the second wave are Zaragoza, Teruel, Toledo and Valladolid, with about one death for every thousand susceptible people. The worst province of the third wave happens to be Soria, which had the worst first wave, and therefore had less susceptibility, but still registers many deaths since Christmas.

What happens to Madrid after making these adjustments? After adjusting for age, it is the fifth province with the worst mortality in Spain. And it had the third worst figure in the first wave. His second and third waves have not been abnormally bad, not even after correcting for how much population he was still susceptible to infection. In the second wave 16 appears (out of 52) and in the third, like 24, close to the middle.

Again, the complexity of making judgments just by looking at these numbers becomes apparent. “It is a multi-causal phenomenon that will have to be progressively unraveled, no easy conclusions can be drawn,” says Clara Prats. “It is normal for an epidemic of this type to show this variability, because it is a complex system,” he continues. No two regions are the same and their characteristics make them more or less vulnerable. We have mentioned two variables – age and susceptible – but it is easy to think of many more: the climate, the population density, the habits of the people, the measures imposed, their compliance. “There are even phenomena of chance that intervene”, explains Alberto García-Basteiro. “For example, in certain outbreaks the origin could be in people who we do not know why they are great disseminators and who were in the right place and time for the spread,” he says in relation to the ‘K factor’ of the virus.

But are the case, admissions and ICU data different?

So far we have talked about death figures, because it is the most homogeneous metric to measure the impact of the virus, as well as the most serious. In addition, there are not many surprises if the cases or the admissions are observed: these are figures, in general, proportional to the number of deaths.

The following graph shows the relationship of confirmed cases in each province and their figures for hospitalizations, ICU admissions and deaths. (We only considered the second and third waves because in the first wave the detection of cases was very bad; in addition, we have again adjusted the rates of admissions and deaths for the age structure).

Mobility is not so different

The mobility of people – our trips to work, to bars or anywhere – has been related to more contacts and a greater risk of contagion. But, do differences in mobility serve to explain mortality by region in Spain?

The data from Google (based on the geolocation of millions of mobile phones) they allow measuring the level of displacement in each province in relation to its normality prepandemic. And no big differences are observed. As you can see from the charts, the movements have gone up and down at the same time. In the first confinement all the provinces were closed; then their mobility has fluctuated (especially in August, when the big cities were emptied), but almost always quite synchronously.

These data also make it possible to distinguish between trips to work and those that are for leisure. In Madrid, the activity in bars or shops stands out in places like Barcelona or Ávila. It is something to be expected because those provinces imposed more restrictions on trade. At the same time, if we look at mobility for work, it is Madrid, Barcelona or Valencia that stand out for having reduced their activity more than the rest of Spain, probably due to the type of jobs and because teleworking is more common.

A related variable, more difficult to measure, is connectivity. “It is not by chance that the islands and peripheral communities such as Galicia or Cantabria are the ones that present the best data, nor that 70% of the dynamics of the first wave can be explained in terms of connectivity with Madrid,” explains Saúl Áres.

These mobility data also remind us that there will be many other circumstances that are common to much of Spain. For example, Madrid, Barcelona and Valencia The three stand out because they have reduced their mobility less than other large European cities. “Madrid has had more lax measures in terms of hospitality, but perhaps there are still many other similar and comparable to Barcelona (mobility and density in metro or bus, group restriction …)”, reminded us María Lahuerta, epidemiologist at the University of Columbia. And the measures go as far as they go: “Despite the harsh restrictions that Catalonia placed on the hotel industry [en Navidades]We know that on those dates the meetings were held a lot as a family, which could also explain that during that wave the restrictions did not help so much ”.

Disclaimer: If you need to update/edit/remove this news or article then please contact our support team Learn more