To get an understanding of what to expect in terms of hospitalizations in the short and medium term, we compared hospital capacity and the epidemiological situation in different European countries. This allowed us to estimate how quickly hospitals were filling up everywhere and where international assistance might need to be mobilized.

To express hospital and healthcare capacity, we looked at four parameters: (1) the number of hospital beds in each country, both in total and the specific number of beds for acute and intensive care; (2) the number of doctors and healthcare providers; (3) the relative number of healthcare providers and the relative number of intensive care beds in relation to the total population, and (4) we also compared national healthcare expenditures (as a percentage of the gross national product) for the different countries in question.

The situation in Italy on March 11, 2020, served as a reference point for a healthcare system on the brink of overflowing. At that time, the country had 10,590 active COVID-19 cases. Since the number of cases, especially during the first wave, was not a good approximation of the epidemic's size and was not very reliable due to significant differences in test availability and strategies between different countries, we also looked at the pandemic's intensity based on mortality rates. On average in Europe, there are 11.5 intensive care beds per 100,000 inhabitants. Italy, with 12.5 beds per 100,000 inhabitants, surpasses the European average. The number of doctors in Italy is also above average, with about 400 per 100,000 inhabitants.

In just two weeks (between March 11 and 25), according to our calculations, the pressure on Italian hospitals increased eightfold. Italy was closely followed by Spain. The Netherlands ranked third, with hospital pressure about 2 to 3 times higher than in Italy on March 11. However, the following countries, namely France, Switzerland, the UK, Belgium, Denmark, Luxembourg, and Sweden, also experienced high pressure on their healthcare systems during that period.

This way of 'scoring' the pressure on hospitals is based on a number of assumptions and depends on the accuracy of certain data. It was not always easy to find accurate figures regarding the number of staff members or hospital beds, especially in the case of intensive care beds. Reports from recent years sometimes provide different figures, and for all European countries, except for Bulgaria, Ireland, Poland, and Romania, the number of beds has systematically decreased over the past 10 years.

The pressure on hospitals is also based on the basic capacity, while we know that exceptional measures have been taken in many places to temporarily increase hospital capacity—such as postponing non-urgent care and making additional beds and staff available. At the same time, the epidemic itself also led to a lot of healthcare worker absenteeism, a factor that was difficult to account for. The increased pressure applies to how hospitals would be overwhelmed by the pandemic in "normal" circumstances, and those results are very clear: the pressure on hospitals would quickly exceed capacity by a significant margin in almost all European countries.

How long do COVID-19 patients stay hospitalized?

In the early weeks of the pandemic, we primarily relied on data from Wuhan to make estimates about the consequences of the increasing number of cases in our country. In parallel, it is important to compare these figures with local data.

The time interval between the first symptoms and hospital admission and how long an individual with COVID-19 typically stays in the hospital is essential for a better understanding of how many hospital beds (both in regular wards and in intensive care) are needed and to assess the situation in local hospitals with the right nuance and context. The time span between the first symptoms and a potential death is also important for calculating the case and infection fatality ratio.

In Belgium, hospitalizations are recorded for all admissions related to COVID-19; i.e. both admissions due to COVID-19 and admissions for other reasons where the patient tested positive. Data is collected through two online surveys: one with information upon admission and another at the time of discharge.

As of June 12, 2020, our country had 14,618 hospitalized patients with COVID-19. In our study, we initially only considered patients who were discharged from the hospital by the end of that period. Patients who were still in the hospital on June 12 were not included.

When broken down into four age groups, this included 258 children and young people up to the age of 19, 4,338 adults in the working-age population (20 to 59 years), 5,480 individuals in their sixties and seventies, and 4,542 individuals aged eighty and above. A significant portion of hospitalized individuals in their sixties lived in nursing homes: nearly 12% of those aged 60 to 79 and over 35% of those aged 80 and above. This is respectively ten times and twice the actual percentage of nursing home residents in those age categories.

Foreign studies report averages of three to ten days between the onset of the first symptoms and hospital admission. In Belgium, this average is 5.74 days, but it varies between three and ten days depending on the age group. For the working-age population, the time span between the onset of the first symptoms and hospital admission is the longest. The period is also two days longer for nursing home residents compared to patients from the same age group. The time between the onset of symptoms and diagnosis is very similar because during the first wave, the COVID-19 diagnosis was often confirmed with a PCR test upon hospital admission.

The duration of hospitalization for COVID-19 patients needs to be considered separately for those who recover and those who ultimately die. Patients who recover typically stay in the hospital for about five days (young people) to around sixteen days (elderly). Patients who die typically spent about six days (elderly) and twelve days (working-age population) in the hospital. For recovering patients, the older they are, the longer they stay in the hospital. Men also require more time to recover than women. On the other hand, nursing home residents leave the hospital more quickly than their peers who still live at home, probably because they can usually rely on the appropriate care and support for further recovery outside the hospital.

For patients who die, the length of hospital stay—and thus until death—is shorter for the elderly than for patients in the working-age population. In this case, the duration of stay is longer for nursing home residents than for their peers (possibly because they are transferred to a hospital earlier, but this is speculative). The same applies to the duration of stay in intensive care.

Between the beginning and the end of the first wave, the average length of hospital stays slightly decreased. This observation holds true even after we corrected for potential bias due to the data cutoff on June 12, which caused admissions with longer durations to be excluded. What could explain a shorter hospital stay? One possibility is that experience and better treatments result in improved care and, consequently, a shorter hospital stay. But it could also be because the overall profile of the average patient changes, and doctors, for example, see more infected individuals without underlying conditions. We know that the latter is a determining factor in the course of COVID-19 and, therefore, a potential hospital stay. In the absence of data on underlying conditions, we could not take this into account, but it would of course be very useful information for future studies.

Regarding the length of stay in intensive care units, Belgium appears to perform better than other countries with 3.8 days for young patients, 6.4 days for the working-age population, and 7.6 days for seniors. In China, a typical length of stay is eight days in the intensive care unit. Studies outside China mostly report around seven days, while in the UK, it goes up to about 12 days. However, these figures are difficult to directly compare because the criteria for hospital admission and discharge, especially for intensive care units, differ across countries.

What is the risk for healthcare workers?

Healthcare workers have been on the frontlines since the beginning of the pandemic. Especially in the early stages, there was a lot of uncertainty about the risks they faced, even with appropriate protective equipment. Some studies found no increased risk of infection among healthcare workers, while others did.

To optimize the approach and procedures in our hospitals, we investigated, based on local data, whether healthcare workers faced specific increased risks of contracting a COVID-19 infection. For this purpose, we collected data from all healthcare workers in the Antwerp Hospital Network, a public hospital with multiple locations and a total capacity of 2,500 hospital beds.

In June 2020, all 6,838 employees of the hospital network, including nurses, doctors, paramedics, administrative, and maintenance staff, were asked to fill out a survey about the potential risks of exposure to SARS-CoV-2 that they had experienced since March 1, 2020. The questionnaire inquired about the time someone spent on COVID-19 units, as well as about high-risk contacts in their private lives or through colleagues; thus, all risk contacts, with or without personal protective equipment. A blood test was conducted for everyone who completed the survey to detect antibodies. Data from the survey were supplemented with information from the staff database (date of birth, age, place of residence, and employment) and then anonymized before being passed on to researchers at the University of Antwerp.

In total, more than three-quarters of the hospital staff participated, which included more than 5,000 individuals aged between 17 and 73, with four times more women than men. Despite the significant gender imbalance, the sample is representative of the entire staff of the Antwerp Hospital Network, which has approximately the same gender ratio. Blood analysis revealed that 397 staff members had antibodies against the coronavirus: 7.0% of men and 7.7% of women. When considering only those staff members with direct patient contact, this figure rises to 8.7%. This result is consistent with that of a sample taken around the same period by Sciensano, the Belgian Institute for Public Health, which found antibodies in 9.4% of 850 healthcare workers from various Belgian hospitals. In the general population during the same period, the figure was slightly lower at 6.3% (but the difference from the overall staff figure of 7.6%, which includes both care and administrative staff, was not statistically significant).

There were more than twice as many healthcare workers with antibodies in the group that had worked in a COVID-19 unit (a unit for suspected or confirmed cases, including intensive care) compared to those who had not worked in such a unit. The longer healthcare workers had been on a COVID-19 unit, the greater the chance that antibodies were found in their blood.

Although there were no differences between hospital sites, there were relatively more staff members with antibodies in two chronic care units (rehabilitation and geriatric patients with dementia), while the percentage of staff with antibodies in the children's unit was significantly lower.

Aantibodies were found most often for nursing staff (1 in 10), followed by medical (1 in 15) and paramedical (about 1 in 17) staff. Only 2.9% of administrative staff had SARS-CoV-2 antibodies in their blood.

What do all these numbers tell us about the risk that healthcare workers face in their profession? The interpretation is not straightforward. First, the timing of the serological test plays a significant role. When these blood samples were collected, Belgium had just come out of a lockdown. Contacts outside the hospital had been very limited. However, as the epidemic progressed and measures were relaxed, this could have changed entirely. Furthermore, it is essential to clearly define who among healthcare workers is genuinely on the so-called frontlines, that is, who was actively working on units with COVID-19 patients for an extended period. Finally, the conclusion also depends on the test's sensitivity; as long as there is uncertainty about this, it remains somewhat speculative.

Even with the right protective equipment, it is difficult to completely reduce the risk of transmission from patient to healthcare worker. The reason that the risk of infection is not higher for intensive care than for "regular" COVID-19 units could be related to better ventilation or a higher number of healthcare workers per bed. Non-protected contacts among colleagues might also play a role.

Antibodies were more frequently found in younger healthcare workers (in their twenties) than in older colleagues, a difference that mirrors the situation in the general population in Antwerp. Perhaps, they had more exposure to the virus outside of work. Of course, it could also be that more experienced healthcare workers followed protective procedures more rigorously.

The results were used to raise awareness among healthcare workers about potential risks and the importance of testing and contact tracing.

The second wave was no surprise

During the autumn of 2020, hospitals were once again overwhelmed by a second wave of patients. This was foreseen in all our models long in advance, but a policy response was delayed again and again.

Our cliquets diagrams for the number of hospital admissions and the rate at which they increased clearly visualize the relationship between the measures taken in August 2020 and the regained control over the pandemic. The cliquets diagram illustrates how severe an outbreak is in terms of the projected number of required beds and the impending pressure on hospitals. In the safe zone, there is a limited number of new admissions, and that number either decreases or increases only slightly. In such a scenario, only a limited number of people with COVID-19 end up in intensive care. In the next zone, we need to be more cautious. Eventually, we reach "high impact" and "no-go" zones, where we greatly exceed the normal hospital capacity.

The surge capacity, or disaster plan, to expand hospital capacity as needed, was also based on this. In phase 0, 15% of intensive care beds are reserved for COVID-19 patients, and in phases 1A and 1B, this number increases to a quarter and half of all beds, respectively. In phases 2A and 2B, 60% of the existing capacity is reserved for COVID-19 patients, and up to 800 additional beds are created.

At the end of September and the beginning of October, we saw a significant increase in the number of new hospital admissions for COVID-19, but measures were not implemented until we were already in the danger zone. Although our models cannot prove causality, it is clear that taking measures early, in other words, before entering a "high impact" zone, keeps the pressure on hospitals under control. This was successful during the summer wave, but in the autumn, the delayed response had severe consequences for hospital occupancy.

In Belgium, we are fortunate to have a relatively large number of intensive care beds, which may have given us a false sense of security, along with the fact that our testing capacity was well-established. Our policymakers were preoccupied forming a new government, and a large autumn wave was met with skepticism by some in the media.