COVID-19 World Review, Part 4: Hospitalizations and Intensive Care

This is the fourth of my review reports on COVID-19 worldwide as at August 20^th 2021. This report will cover hospital bed and intensive care unit (ICU) occupancy by COVID patients. These numbers are important, as they determine how fast new case levels can be allowed to rise without over-extending finite health care system resources.

Countries Reporting

In contrast to cases and deaths, only a small minority of countries are reporting hospital and intensive care occupancy by COVID patients. Here’s the list of 30 countries reporting the current number of COVID patients in hospital per million population, in order:

Of these countries, only the USA, Israel and Canada are outside Europe. I will, therefore, only show spaghetti graphs for the four groups into which I divide the 46 European countries, and for North America. For Israel, I’ll show graphs covering Israel only.

Here’s the equivalent graph for intensive care unit occupancy per million population:

The members of the two lists are slightly different. Germany, Romania and Algeria are providing ICU data, but not hospital occupancy. So, I’ll do the same graphs for Algeria as I do for Israel. I’ll do the same for Finland, which is the only one in its group reporting ICU data. Lithuania, Iceland, Croatia, Latvia, Hungary, Poland, Slovakia and Norway are all providing hospital data, but no ICU data.

Hospital Occupancy Graphs

Here are the spaghetti graphs of hospital occupancy per million by COVID patients over the period of the epidemic:

The population of Israel is about 9.2 million, so that peak just shy of 2,500 patients represents around 270 patients per million.

Hospital occupancy as a percentage of beds available

For those countries for which I have an estimate of the total number of hospital beds (which includes all 30 countries reporting hospitalization data), I can convert the current hospital occupancies per million to a percentage of available hospital beds which are occupied by COVID patients. The result is as follows:

For comparison, here are the peak values of this metric from earlier in the epidemic for the “Europe 14” group of countries:

So, none of those current percentages seem to give great cause for concern, when compared with the earlier peaks; at least in Europe.

Hospital occupancy per case

It’s clear from the above spaghetti graphs that, in most of the countries, hospital occupancy by COVID patients now is considerably lower than it was during earlier peaks of the epidemic. This is to be expected, because a substantial proportion of the population in most of these countries has now been vaccinated, and the vaccines are reputed to cut COVID hospitalizations. To try to see how much effect this has had, I have plotted some graphs of hospital occupancy per new case. I found two difficulties with doing this. The first is how far back to look in the new case numbers to compare with; I arbitrarily picked 7 days back, one third of the 21 days offset I use when comparing cases with deaths. The second is that hospital occupancy depends not only on the number of cases requiring hospitalization, but also on the “dwell time” that each patient spends in the hospital. Since I did not have a figure for the average dwell time, and suspected that it might vary quite significantly from country to country, I decided not to try to allow for this. This is why the percentages shown on these graphs are many times higher than the proportion of cases which require hospitalization.

Note the big spikes around June 2021 in (at least) France, Poland, Bulgaria, Canada and Israel. It looks as if the arrival of the delta variant in a country causes a spike in hospitalizations per new case. However, in each case, the spike has now subsided considerably. That would suggest that a new variant is initially more virulent than the variants it is competing with; but as time goes on and it takes over as the primary variant in a country, that virulence decreases.

Here are the lists of hospital occupancies per new case (7-day offset) for the European countries:

There is a huge variability between different countries, though the figures for Western European countries are far smoother than for those further east. I can’t tell whether the high outliers like Poland, Hungary and Bulgaria are caused by more hospitalizations needed per case for the new variant, or a longer “dwell time,” or a combination of both.

Of the non-European countries, Canada came in at 40%, Israel at 22% and the USA at 67%.

How much difference have the vaccines made to hospital occupancy?

Comparing the hospital occupancy per new case now against last February, when the vaccination programs were only just really getting under way, there does seem to have been a decrease in most countries. I chose February 20^th as my baseline. Not only because it is exactly 6 months prior to the cut-off date of my data. But also because it represented a relatively flat period in hospital occupancy per case for most of the countries reporting.  

But the magnitude of the decrease seems to vary a lot from one country to another. I calculated the ratio of occupancy per new case now to the occupancy per new case six months ago for each reporting country in these groups, then grouped the countries by this ratio, with the following results:

·       Under 10%: Iceland.

·       10% to 19%: Denmark, Ireland, Norway, UK.

·       20% to 29%: Portugal, Spain.

·       30% to 39%: Belgium, Croatia, Czechia, Finland, France, Italy.

·       40% to 49%: Austria, Lithuania, Netherlands, Slovakia, Slovenia, Switzerland.

·       50% to 66%: Bulgaria, Cyprus, Estonia, Hungary, Latvia, Sweden.

·       67% or more: Luxembourg, Poland.

Of the non-European countries, Canada came in at 50%, Israel at 58% and the USA at 92%.

These figures seem to indicate that in most places in Europe, the vaccines have helped to cut hospital occupancy per new case by a factor of between 1.5 and 5 compared with what it would have been without them.  The UK has done particularly well, with hospital occupancy per new case now being only about a sixth of what it was six months ago.

For comparison, here’s the list of percentages of the population fully vaccinated in Europe:

It does look as if there is at least some positive correlation between vaccination levels and lowering in hospital occupancy per new case since February. Though Norway is an outlier, with a good improvement in hospital occupancy per case, but a relatively low vaccination level.

ICU Occupancy Graphs

To intensive care. Here are the spaghetti graphs of ICU occupancy by COVID patients:

There is, as you would expect, a lot in common between the graphs of hospital and ICU occupancy in most, if not all, countries. It would be interesting, if I had the time, to look at how the proportion of COVID patients who, once admitted to hospital, need an ICU bed, varies by country. But no. I’ll pass on to…

ICU occupancy as a percentage of beds available

For those countries for which I have an estimate of the total number of ICU beds, I can convert the current occupancies per million to a percentage of available ICU beds which are occupied by COVID patients. The result is as follows:

For comparison, here are the peak values of this metric for my “Europe 14” group of countries:

The Portuguese, Dutch and Spaniards must have been building a lot of emergency ICUs! But, as with hospital occupancy, ICU occupancy by COVID patients now is only a fraction of what it was back at the peaks – even in Spain, France and the UK; although Spain has not reported any ICU figures in over a month.

ICU occupancy per case

For ICU occupancy per case, I picked an offset of 14 days between the new cases and ICU occupancies which I compare to each other. Here are the results:

Again, we see big spikes in June/July 2021 in certain countries, notably: Germany, France, Austria, Romania, Bulgaria, the USA, Canada and Israel. Very much the same countries, that had spikes in the hospitalizations! In all these countries, the spike is now on the way down.

Here are the lists of ICU occupancies per new case (14-day offset) for the European countries reporting (except Finland):

Algeria came in at 2.8%, Canada at 45.9%, Finland at 3.4%, Israel at 22% and the USA at 29%.

I did the same exercise as before, working out the ratio of ICU occupancy per new case now to ICU occupancy per new case six months ago, and then grouping the countries. The ICU occupancies are far more volatile than the hospital ones, so the results are a lot less certain. But here they are:

·       Under 20%: Denmark.

·       20% to 24%: Ireland, UK.

·       25% to 39%: there were none.

·       40% to 49%: Algeria, Finland, Italy, Netherlands, Portugal.

·       50% to 59%: Belgium, Czechia, France, Spain.

·       60% to 69%: Austria, Germany, Romania, Slovenia, Sweden.

·       70% to 79%: Cyprus, Estonia.

·       80% to 99%: there were none.

·       100% to 119%: Bulgaria, Israel, Luxembourg, Switzerland.

·       120% to 199%: there were none.

·       200% or more: Canada, USA.

This suggests that, in Europe, vaccinations may have had some effect on lowering COVID ICU occupancy; though it is smaller, and far less clear, than the effect on hospital occupancy. In 16 out of the 25 countries reporting, ICU occupancy per new case has been cut over six months by a factor of between 1.25 and 2.5. Three countries (Denmark, Ireland, the UK) have done substantially better than this; while the remaining six, including the USA and Canada, now have similar or higher ICU occupancy per new case than they did six months ago.

It’s worth noting that Algeria has done relatively few vaccinations compared with the European countries (less than 2% fully vaccinated). And yet, it has been one of the more successful countries in terms of lowering ICU occupancy per case. This means that vaccines cannot be the only factor contributing to this lowering.

To sum up

Neither current hospital bed nor ICU occupancy percentages (among 30 and 25 countries reporting respectively) seem to give great cause for concern, at least in Europe, when compared with the peak values from earlier in the epidemic.

In most of the reporting countries, hospital occupancy per new case (with a 7-day offset) has been cut since February 2021 by a factor of between 1.5 and 5. There is at least some positive correlation between vaccination levels and lowering of hospital occupancy per new case since February.

In 16 of the 25 countries reporting, ICU occupancy per new case (with a 14-day offset) has been cut since February 2021 by a factor of between 1.25 and 1.5. Three countries (Denmark, Ireland, the UK) have seen the ratio fall by a factor of 4 or more. The remaining six, including Canada and the USA, now have similar or higher ICU occupancy per new case than they did six months ago. The effect of vaccines on ICU occupancy is far less clear than it was for hospital occupancy.