18+ 
A new epidemiological index for the COVID-19 pandemic
16.04.2021
The coronavirus epidemic has so far caused thousands of victims, and the economic damages are added to the grief and the spiritual suffering that citizens all over the world have already felt. Italy is no exception in all this and, indeed, our beloved country continues to find itself in a state of exception that seems to have no end while the “vaccination campaign” limps in spite of bombastic proclamations.
The fight against the pandemic develops along various directions, one of which is the statistical modelling of the development of the disease, in order to understand which scenarios await us and if the containment measures adopted are truly effective.
There have been (and still are now) many analysts, journalists and bloggers who have questioned the validity of the Western approach but also the scientific community is not completely homogeneous in its judgments on certain measures, for example the generalized use of quarantine that places in isolation the healthy as well as the sick. John Joannidis of Stanford University [1] doubts about the real effectiveness of this generalized option, but also Italy tried to explore alternative directions.
In a recent article published in the prestigious scientific magazine Nature [2], the research group of Professor Mariano Bizzarri of the University "La Sapienza" of Rome, proposes the use of a new epidemiological index RI, finer and more predictive than the one so far used and that greatly conditions the choices of governments and local administrations.
Professor Bizzarri, whose impressive resume [3] ranges from oncology to space medicine and beyond, has agreed to answer our questions about this particular research of his. The topic is not easy but we tried to make it accessible to readers without impoverishing it of its scientific depth.
1) Could you remind our readers what is meant by statistical model and why it is necessary to use statistics to study an epidemic?
A) The statistical model allows to define a relationship - expressed in mathematical terms - between the quantitative parameters of a phenomenon that are the most suitable for describing that phenomenon and its dynamics, that is, the evolution over time. This approach makes it possible to a) identify the factors that can change the parameters and therefore the evolution of the phenomenon; b) predict the subsequent stages that the process in question takes, in our case the epidemic.
2) How many models have been proposed for the study of this epidemic and which, among them, proved to be the best?
A) Many models have been proposed but in our opinion - mine and colleagues with whom we have given substance to the study then published in Nature Scientific Report [2] - but none captured the salient parameters of the COVID epidemic.
3) Why does the spatial distribution of the pandemic in Italy make it difficult to determine its trend over time?
A) In reality, the unequal distribution is only apparent and depends - precisely - on the wrong choice of the descriptive parameters of the model. In fact, if we consider the evolution over time of the relationship between hospitalized in intensive care (Fig. 1), it can be observed that all regions have a similar trend. We thus understood that the main parameter for understanding the evolution of the epidemic, and the main index of its real severity, is precisely the number of hospitalizations in intensive care. Other parameters, if not included in a complex model, are not immediately interpretable and can be used incorrectly or exploited.
Fig.1 – Trend in the number of hospitalized in intensive care in Italy without Lombardy (left) and trend in Lombardy alone (right) up to May 2020
4) What is the R index that our media often tell us about and that conditions the choices of governments in managing the epidemic?
A) The R index (i.e. the basic reproduction number) - known as R0 - indicates the potential transmissibility of an infectious disease. It expresses the number of new symptomatic cases “produced” by each new single infected case and therefore constitutes the expected number of new infections originating from a single individual during his entire period of infectivity, in a population entirely susceptible at the onset of an epidemic or in contexts in which no measures have been taken to limit the infection. This definition immediately reveals the limits of the parameter itself given that 1) it assumes that the population is ALL equally sensitive; 2) does not take into account the limitations adopted (quarantine, masks, vaccinations etc.); 3) has a “static” value, i.e. it does not provide information on the speed with which an epidemic will spread through a population. The index has no predictive value, but is only retrospective and, above all, it does not help to understand the evolution of the severity of the disease. In fact, the index has been universally criticized as inadequate and numerous scientific publications attest to this. Nonetheless, it has been uncritically (and irresponsibly) used by politics and health administrations as the main parameter for monitoring the epidemic. The inconsistency in the name and definition of the R0 parameter was potentially a cause of misunderstanding of its meaning.
5) In the article you and your team propose a new and different RI index. Can you explain the formula that expresses it and why do you think the new index would be more suitable for the purpose?
A) The index that we propose is a “dynamic” index precisely because it relates the speed with which the number of infected varies with the speed with which the number of healed changes. This relationship (which is a relationship between derivatives) graphically describes a scissor trend (Fig. 2).
Fig. 2 - Trend in the number of infected and healed in Italy up to September 2020.
When there is a decussation between the curves of the healed (who grow) and that of the infected (who decrease), it is then certain that the epidemic is waning.
6) So, can we say that the adoption of this new RI index would therefore allow us to improve the management of the epidemic, at least in Italy?
A) Absolutely yes, because it allows to understand the dynamics of the process and to identify the moment in which the curve deflects. The simple observation of the infectivity rate (number of infected compared to the number of swabs) is of no use, also because the reliability of the tests used is burdened by a large number of false positives and negatives that make the parameter very unreliable. Incidentally, this is why the contagiousness rate tends to have continuous daily variations and gives the illusion that once the epidemic increases and another decreases in just 24 hours. Obviously, this is a blatantly false figure. Yet it is on these inaccurate data that the media narrative of the epidemic was built.
7) Does the model you use also allow us to identify with some confidence the temporal beginning of the epidemic in our country?
A) Yes, the retroactive calculation made it possible to place the beginning of the pandemic in Italy in September 2019.
8) If this coronavirus is as infectious as we are told and also was in Italy since the second half of 2019, why did the terrible scenario of Bergamo and Brescia only occur in March 2020? Why haven't there been similar emergencies in other parts of Italy and maybe even earlier?
A) This remains an enigma. Why Lombardy has been hit so hard deserves a specific investigation and evaluation of concomitant factors that have made the elderly population of that area so susceptible to the most severe form of the disease.
9) Based on what you write, it would seem that the epidemic in Italy reached its peak in October 2020. Shouldn't it have gradually run out? Why, in this our April 2021, are we still living in a state of emergency, according to an end of the world scenario?
A) There are no models that can predict whether an epidemic - once a first phase is over - can then reactivate. It is common for this to happen with influenza viruses (which, among other things, like COVID, mutate quickly), but it is not mandatory. In reality, the second wave was not predictable: it could happen, as it did, but it was not obligatory for it to happen. As for the “apocalyptic” scenario, I believe that it must be resized by paying attention to the true parameter of severity: the number of hospitalizations in intensive care. However, I believe that a modern country with over 60 million inhabitants can manage an average occupation of about 3000 patients in intensive care without abandoning itself to the blind desperation and media terrorism that we have to endure every day.
10) Were there any reactions from politicians and scientists to your ideas? Or has your work been virtually ignored?
A) Unfortunately - except from the appreciation of some colleagues, including epidemiologist John Ioannidis of Standford University - the health and political authorities have not received anything. However, that doesn't surprise us that much.
[3] Professor Bizzarri's resume is truly impressive (Italian only):
