In view of the official figures released today by the Directorate-General for Health (DGS), which indicate 345 deaths from COVID-19 so far in Portugal, the projections of researchers from the Institute for Health Evaluation and Metrics (IHME), at the University of Washington, point to another 126 deaths over the next three months.
The study also mentions that the mathematical model used for the estimates, which were made for several countries, "demonstrates that, although Portugal isn’t suffering from a total lack of beds, it did not have sufficient intensive care units (ICU) for the demand, with the peak causing a lack of 118 beds on the 3rd of April ". At a European level, the IHME estimates that the daily peak in the death rate due to the COVID-19 pandemic is expected to be reached in the third week of April.
Mathematical models are used to make predictions. In a simplified way (not always, let's face it), they represent reality in a mathematical way. And in the case of an epidemic or pandemic, these models are used to estimate the number of infected people that may exist. For this purpose, variables such as, for example, the speed of contagion are combined, assessed by the number of people that someone infected may infect, the number of new daily cases and/or the age division of the population.
But epidemiologists have models that include other variables. Carlos Antunes is a geographer and engineer. He applied the mathematical models to epidemiology that he uses in projections related to climate change. And the effectiveness of their projections has been surprising, as they are very close to reality.
In the case of the evolution of Covid-19 in Portugal, it indicates that “we are leaving the peak”. The same is to say that we have probably passed it. Taking into account the figures released on Monday by the Directorate-General for Health (DGS), Carlos Antunes assures us that we are now at a stage where we are likely to see the percentage increase of cases decreasing.
“We will have passed the peak. Contrary to what we were calculating with some methodologies, the peak will have occurred between March 31/April 4. Because the trend in the number of new cases is effectively decreasing”, he told press.
The modelled scenarios are made with some degree of certainty for the next five or ten days. An essential period of time for planning by health authorities, who can predict in advance how many more cases they will have to manage.
In Mr. Antunes’ projection, the peak of the new case curve corresponds to the inflection point of the accumulated case curve. In simple terms, this means that from this point on, “growth is no longer exponential”. An initial feature of epidemics and pandemics.
But one must be careful: mathematical models are sensitive to various values. One is the basic daily spread of the infection. A slight fluctuation in this and other values makes the models respond differently. “Of the approximately 87 thousand cases already tested, only about 13.4% have been confirmed. Seventy-one thousand have not been confirmed, but there may be many of those unconfirmed in which the viral load is so low that the test was unable to identify it”, explains Carlos Antunes.
Therefore, caution remains part of mathematical modelling. “It is necessary to be cautious about these cases. The numbers indicate that there is already a passage through the peak and that we are in a phase of decreasing new cases. If that translates to reality, we don’t know yet”, he concludes, because it will take one to two weeks to assess the reality of the current situation.
Although the Minister of Health, at a March 21st press conference, stated that the largest number of cases in one day would be mid-April, the same Minister, Marta Temido, and the Director-General of Health, Graça Freitas, have also stated that the date may be much later, in mid-May. At that time, the director-general of Health added that the peak of COVID-19 should not be reduced to just one day, but that it should be, "a plateau of more or less similar cases over several days".
