ISS, the risk assessment system: how and why it works - ISS (EN)

ISS, 17 November 2020 – What data are used in assessing risks?

Twenty-one indicators were selected for the risk assessment, 16 of which are ‘mandatory’ and 5 optional, which make it possible to evaluate three aspects of interest for risk assessment: the probability of the epidemic to spread, the impact on healthcare systems and regional resilience. The complete list is available in the Health Ministerial Decree (DM) of 30 April 2020.

Why use all these indicators?

Several indicators were chosen from various information inputs because, especially in emergency situations, there is a higher risk that the data may be affected by the overload on healthcare systems and therefore may not be optimally complete and timely. In epidemiology, an analysis is considered to be more grounded when several information sources confirm the same trend (for example, an uptrend in the number of cases).

How are data collected and processed?

The data are submitted by local agencies to the Regional Authorities, which in turn transmit them to the Ministry of Health and to the ISS. Algorithms are applied in the data processing that, when combined, make it possible to assess the risk in every Region on a weekly basis.

Isn’t the system too complex?

The system is necessarily complex in order to include all the aspects linked to the epidemic and to the response of the healthcare systems. The presence of different indicators also makes it possible to limit the negative effects of the system overload on the completeness of the data available.

Are the data on which assessments are made obsolete?

No, the consolidated data from the previous week are analysed every week except for some data, for example the hospital bed occupancy rate, for which a more updated reliable number is available. Considering that the incubation period of SARS-CoV-2 ranges from 5 to 14 days, this update frequency is sufficient to evaluate the trend of the epidemic: a trend analysis enables us to understand “where we’re going”. Moreover, the use of prospective indicators like the Rt, 30-day projections of hospitalization rates, and making the best use of data on new hotspots involving fragile population segments (which are likely to need hospital assistance after a few weeks), enable us to “look ahead” while using the same data.  

Why are the data used the most updated possible?

The acquisition of epidemiological data on infections is affected by a number of delays, some of which are not compressible: in particular, the time between the infective event and the onset of symptoms (incubation period), between the onset of symptoms and the swab test, between the swab test and the positivity result and the time between the confirmed positivity and inclusion in the ISS Integrated Surveillance System. The comprehensive delay between infections and their inclusion in the Surveillance System is evaluated and updated weekly by analysing the stability of the number of cases (symptomatic or hospitalised) recorded each day. This evaluation process is the basis for choosing the most recent date at which various Rt estimates can be considered to be sufficiently stable.  
It should be pointed out that the possible delays in administering tests and swabs due to the increased number of infections, have an impact both on the aggregated number of new positive cases reported daily by the Civil Protection Department and on the data contained in the Integrated Surveillance System.

One of the data most often quoted is the ratio between positive cases and people tested. Is it reliable seeing that single Regions decide to only include the people positive to the molecular swab test or also those positive to other types of tests?

At present, the international definition of a case requires that the SARS-CoV-2 virus infection be confirmed through a molecular test. This is why this definition, for now univocal, applies to the provisions made in the circular letters of the Ministry of Health. Every day, the Ministry of Health coordinates the data collected from the Regions and Autonomous Provinces on the number of swab tests performed and, once the data reliability is assured, transmits them to be published on the Civil Protection Department’s website.

R0, Rt: what are they and how are they calculated?

The basic reproduction number (R0) of an infectious disease is the average number of infections transmitted by every infected individual at the beginning of the epidemic, in a phase in which no specific actions (pharmacological and not) are usually taken to control the infection. R0 therefore represents the transmission potential, or transmissibility, of an uncontrolled infectious disease. The R0 value is a function of the probability that an infected person will transmit the virus through contact, the number of contacts of the infected person and the duration of infectivity. The definition of the net reproduction number (Rt) is the same as that of the R0, except that the Rt is calculated over time. For example, the Rt makes it possible to monitor the effectiveness of the actions taken during an epidemic.

Why does calculating the Rt only on symptomatic or hospitalised cases make it reliable also when contact tracing systems are failing?

The statistical method of calculating the Rt is robust if it is calculated on the number of infections registered according to criteria that are sufficiently stable over time.
Region by region, the criteria used to detect symptomatic cases or to hospitalise the most serious cases are constant and the number of this type of patients is therefore strictly correlated to the transmissibility of the virus.
On the contrary, detecting asymptomatic infections greatly depends on the capacity of prevention departments to carry out screenings and contact-tracing, which may vary in different phases of the epidemic. For example, this capacity typically increases when the total incidence of the disease, and therefore the workload of the healthcare system, decreases. Consequently, in this context, a greater or smaller increase of asymptomatic cases over time does not directly depend on the transmissibility of the virus. This is the reason why the R0 and Rt estimates that we provide do not take into consideration asymptomatic infections.
Therefore, as of February 2020, it was decided to estimate the transmissibility of SARS-COV-2 in the different Italian regions on the basis of the daily curve of symptomatic cases, insofar as they are less affected by the changes made in Italy’s diagnostic testing policies of asymptomatic patients and hospitalised cases, which are taken as the basis to forecast hospital occupancy rates in the subsequent 30 days.

Why are the intensive care and overall medical occupancy rates disclosed but not the number of persons admitted to and discharged from hospital every day?

SARS-CoV-2 is a virus that provokes a disease that may last several weeks, determining long hospitalisation periods both in general medicine and intensive care units. This is why an uncontrolled increase in new cases of infection not only generates an increase in the number of new accesses to said services but also the progressive saturation thereof. Therefore, the hospital bed occupancy rate is a useful indicator of when hospital beds not occupied by COVID-19 patients risk to become insufficient to assure medical care to the population for this and other pathologies.  

Why are all the available data not published in a searchable database?

In our Country, not all data are public and disaggregate so as to assure the application of privacy protection provisions and of the ordinances that regulate the epidemiological surveillance of the disease. We are working concertedly with the competent authorities to develop further ways of accessing data while complying with these provisions.