Check the figures

How is Covid-19-associated mortality calculated?

ISS, 23 November 2020

How are COVID-19 cases recorded in the ISS Surveillance System?
The ISS Surveillance System records the patients positive to the SARS-CoV-2 molecular swab test diagnosed on the Italian territory. The data are validated by the19 Regions and 2 Autonomous Provinces and sent to the ISS Surveillance System which then produces nationwide statistics. In addition to measuring the incidence of the spread of the disease, the System also collects data on the deaths.  

How are COVID-19-associated deaths defined?
In reporting COVID-19 deaths to the Surveillance System, the ISS suggests to follow the indications of the ECDC and of the WHO in identifying the deaths associated with COVID-19. These indications are laid down in the Report on the definition, certification and classification of the causes of death from COVID-19 which contains the instructions on how to define a death as due to COVID-19 and how to fill out the death certificate.

How can I be sure that a death was caused by COVID-19 and not by some other pathology?
The criteria to define a death from COVID-19 are indicated in the aforesaid Report and include:

  • The death of patients definable as microbiologically confirmed cases (through a molecular swab test) of COVID-19
  • The presence of a clinical picture and imaging tests suggestive of a COVID-19 infection
  • The absence of a clear cause of death other than from COVID-19
  • The absence of a period of complete clinical recovery between the onset of the disease and the death.

If the death is caused by an event that cannot be directly correlated to COVID-19, for example a heart attack, even if the patient is positive, how should the death be classified?
Testing positive to Sars-Cov-2 is not enough to consider the death to be due to COVID-19, because it requires the presence of all the aforesaid conditions, including the absence of another clear cause of death. However, it should be specified that the pre-existing pathologies (such as cancer, cardiovascular, kidney or liver diseases, dementia, psychiatric disorders and diabetes) that could have favoured or predisposed a negative course of the infection should not be considered to be clear causes of death different from COVID-19. The COVID-19-related causes of death include the complications or consequences of pre-existing pathologies that could have favoured or predisposed a negative course of the infection in a patient with a clinical picture compatible with COVID-19. More specifically, if a heart attack strikes a cardiopathic patient with a COVID-19-associated pneumonia, it could be inferred that the heart attack is a complication from COVID-19 and therefore the death must be classified as due to COVID-19. If the heart attack occurs in a patient whose clinical picture is not compatible with COVID-19, the death must not be classified as related thereto.

Is the number of deaths in Italy overestimated?
No, in fact it was probably underestimated in the months of March and April. In this period, many patients died without being tested and therefore their data were not entered in the Surveillance System. The estimate contained in the ISS-Istat joint report on excess mortality showed that in the months of March and April the deaths directly or indirectly associated with COVID-19 were almost twice as many as those calculated in the Surveillance System. This underestimation of the number of deaths was in any case greatly reduced and almost eliminated from May till the end of the summer. In the last few months, the Surveillance Systems have been witnessing a new surge in deaths. Soon it will be possible to estimate a possible excess mortality in the autumn and winter months by comparing it with Istat’s mortality data.  

Are therefore all the deaths reported to the Surveillance System caused by COVID-19?
An analysis of the death certificates, jointly performed by the ISS and Istat, showed that COVID-19 is the direct cause of death in 89% of the deaths reported in the Surveillance System, therefore in almost 9 out of 10 deaths surveyed. This analysis included the deaths in the period between March and May and an updated analysis of the data is currently underway.  

Lethality: what is it? Is it a reliable index?
The lethality index indicates the proportion of deaths among the patients suffering from a given pathology. It is expressed as the ratio between the number of deaths and the number of cases diagnosed with a given pathology. In the case of the SARS-CoV-2 infection, a WHO document (Scientific Brief of 4 August 2020) states that the lethality index can be a misleading parameter especially when used to compare data from different Countries. The document states that Countries are difficult to compare for a number of reasons. They can be more or less inclined and capable of detecting all the SARS-CoV-2 infections and reporting all COVID-19-associated deaths. In particular, Countries could use different definitions of cases and testing strategies and count the cases differently (for example, not counting mild untested cases). It should also be pointed out that the patient’s profile (e.g. age, sex, ethnic group and underlying comorbidities) can vary from one Country to another and this could explain part of the differences. In particular, it should be noted that Italy is one of the Countries with the highest aging index and this explains most of the differences with other European Countries. The difference in the lethality index of different Countries can also be partially explained by the delay in reporting deaths, which does not exclude that in other Countries the lethality index could be underestimated. The ISS had pointed out these criticalities linked to the use of the lethality index in an  international publication released in March 2020. Lastly, it should be kept in mind that our Country’s lethality trend over time needs a cautious interpretation. The progressive growth in diagnostic capacity in the course of the epidemic has led to an increase in the asymptomatic or paucisymptomatic cases detected and reported to the Surveillance System. This has translated into a drop in lethality over time that must be taken into account when interpreting the trend.    

So, how can we estimate the impact of COVID-19 on the number of deaths?
One of the most effective strategies to measure the impact of COVID-19 on the number of deaths is to measure excess mortality, meaning thereby the number of deaths (for all causes) outnumbering a Country’s death toll in the previous year. This excess mortality is normally expressed as a percentage (i.e. by what percentage have deaths for all causes increased in the Country from previous years). A recent OECD report states the following: “By  examining  the  total  number  of  deaths  recorded,  many  of  the  different  ways  that countries record deaths from COVID 19 are removed. By comparing overall numbers with the level  of  expected  deaths  in  a  country  based  on  the  same  period  in  previous  years,  excess mortality can provide an indication of the overall impact of COVID 19 –by accounting not only for deaths directly attributed to COVID 19 but also those that may be missed or indirectly linked, such as deaths caused by delayed or foregone treatment due to an overloaded health system.

What is the excess mortality in Italy and in other European Countries compared to previous years?  
The above-mentioned OECD report analysed excess mortality (i.e. the increase in the number of deaths compared to the preceding years) in several European Countries over a 10-week period starting from last March. This report revealed that the total number of deaths recorded in Spain showed a 61% increase from the average number of deaths registered during the same period of the previous 5 years. In the United Kingdom the number of deaths recorded was 56% higher than in previous years. Italy and Belgium recorded a 40% increase. Germany, Denmark and Norway reported 5% more deaths over the 10-week period. The OECD report points out that these data should be interpreted with caution because the first phase of the epidemic varied from Country to Country and therefore will require a longer period of observation in order to interpret them.

What the ISS publishes on COVID-19 deaths
Since the beginning of the epidemic, the ISS has published and continues to release reports and estimates on COVID-19 deaths. It is essential to analyse these publications in the light of the aspects listed above.

The above publications go to add on to the scientific publications on COVID-19 deaths authored by ISS researchers.