These notes are strictly personal but express ideas which are shared by a vast community of researchers, including epidemiologists, clinicians, immunologists and lab scientists. This is meant to be a working document.

Possible scenarios for the exit phase of the pandemic period

The papers produced by Neil Ferguson’s group at Imperial College have inspired the policies of various countries. One of the most recentproposed that the strategy of bland mitigation involved a high health impact and that a more vigorous strategy of suppression was necessary (1); this strategy was then implemented by many countries. The same study also proposed a strategy of lifting each of the measures at the start of the exit from the lock-down phase. This third strategy of “stop and go” was proposed based on the need of intensive care beds and the predicted number of deceased. In essence, this strategy proposes the implementation of restrictive measures for a certain period of time, and then the lifting of some of the same until the saturation of beds and number of deaths reach unacceptable levels; at that point the measures would be implemented again. The key open questions of this strategy are: which elements of suppression should be lifted and what the duration of each of the two phases should be (stop-and-go).

A more recent paper (2) uses a method of “back calculation” to estimate the impact of the measures taken weeks before, on the basis of the reported number of deaths. It suggests that the containment measures adopted in Italy have been effective, with a value of R0 (the number of infections caused by an infected individual) close to 1 around March 11th. Although the implemented measures vary from country to country, Italy is one of the countries (amongst those analysed) which has implemented them all, that is case isolation at home, social distancing, prohibition of public events, school closures and nationwide lockdown (home isolation for all, all travel banned, only workers of essential services allowed to move). Furthermore, the paper estimates that there have been around 6 million people infected in Italy in total as of March 28th, and that 38.000 deaths have been avoided thanks to the measures taken by the government (but with a wide confidence interval).

As of today, we are observing a progressive decrease in the daily number of cases. It is clear that the effectiveness of the measures that have been put in place and the current deceleration of the epidemic growth curve do not allow for the relaxation of the measures at this time. To this end, modellers are also working on understanding how the relaxing of the different measures of containment (for example social distancing and school closures) can influence the trend and therefore the number of deaths and ICU bed demand. In these estimates, the distribution by age and geographical area are also considered. It is currently too premature to draw conclusions on any practical implications, but this could be the starting point for decisions to be taken. Assuming a large positive effect of the national lock-down, an aggressive action to actively identify and interrupt local transmission has to be planned, overriding the recommendation of limiting viral detection only to symptomatic, severe cases.

The data available for our country clearly identifies the elderly population, with chronic diseases, as being the group most vulnerable to COVID-19. Currently in less affected areas in Southern Italy, outbreaks occur in nursing homes for elderly and disabled. The exit strategy must involve the implementation of a programme of active surveillance that also includes remote monitoring of the health conditions of this population. With every proposed strategy particular attention should be paid to the issue of healthy inequality and access to healthcare.

Scientific data availability from Italy

At present, the national information system of infected cases is managed by Istituto Superiore di Sanita’ and Protezione Civile on the basis of the reports made by the regional health services. We believe that the data collected for monitoring the pandemic in Italy should be shared with the research community. We understand that calls for access to data are coming from different parties, and this would greatly upsurge both epidemiological and clinical research and timely contribute to the advancement of science which should guide the response. Amongst the numerous opportunities for research is the need to try and understand why the lethality of the disease appears to be so variable across European areas with the extreme cases of Lombardy (high) and of Germany (low). Several hypotheses have been put forward, none of which is convincing in its own right. In particular we ought to try and understand the role played by the different organisations and primary care resources and the regional management of the epidemic.

A rational use of tests based on the viral RNA detection

There is a broad consensus that the investigation of viral infections secondary to SARS-COV-2 should be aimed at (a) confirming clinical diagnosis (symptomatic cases), (b) confirming recovery, (c) contact tracing and isolation to interrupt the chain of transmission, (d) monitoring the health of those working in environments where they are potentially exposed to infection and could in turn infect other people.

At the peak of a pandemic the priority is to diagnose the symptomatic cases and the confirmation of those in recovery. To contain the spread, it is necessary to find each individual contact and isolate them. The general lock-down will be effective on the currently unidentified infected contacts, but not on the household transmission, so at least the cohabitees of the positive cases ought to be tested. Those working in essential services, in contact with many people, should be periodically tested and isolated if positive. If they test positive their household contacts need to be monitored in the same way as those who are ill. It is a priority to take care of the more vulnerable groups of the population such as care home residents: suspected cases should be isolated as soon as possible (even whilst waiting for the result of the test), hence the importance of monitoring of this group using apps on smartphones (managed by carers) - which will be discussed further on.

The search for those infected in the general population, such as screening through swab testing, is not very efficient as infection does not spread by chance but follows traceable routes of transmission. Testing should be based on prioritisation criteria as follows:

Immediate priority ought to be given to swab all health workers (particularly hospital staff) who have a high infection risk, and then other key workers who are in contact with the public. All the contacts of those who test positive should be tested as well. A negative test for viral RNA is valid only for the moment it is performed, therefore it is also important to monitor early signs and symptoms to isolate the cases.

Accuracy and predictive values must be monitored. It is essential to understand the sensitivity, specificity and predictive values of the laboratory tests. When the prevalence of the infection in the population diminishes (due to the lock-down effect) the predictive values will change. Low sensitivity indicates many false negatives, that is, people reassured by the results but in reality still contagious. Low specificity implies many false positives, and therefore it would be a waste of resources to isolate them and trace their contacts. Low specificity translates into a low positive predictive value (which could be as low as 50%, i.e. if 100 people tested positive only 50 would actually be infected with the virus). Work is being carried out to improve the efficiency of testing of the virus (high-throughput). An appeal signed by almost 300 researchers in Italy has made available the use of laboratories, equipment and staff to undertake wider testing. These are precious resources that ought to be used immediately. With regard to the tests it should be made clear that the test for viral RNA allows for the identification of those who are contagious and the positivity lasts for a maximum of 2-3 weeks, after which the test becomes negative. Commercial rapid tests must undergo validation before extended use.

Early identification of hotspots and use of technology

The search for those infected cannot be carried out by population screening, but in the early stages for hotspot detection, or during the lifting phase of the suppression measures. For example it may be considered during the partial relaxation of the lock-down. In the preparation for a second peak in the autumn active identification of the cases and contact tracing must be reinforced as they will be crucial. With regards to the CFCT methods (case finding contact tracing) there have been innovative approaches also in Italy that should be harmonised and made widely available. For example, the use of an app for the early detection of symptoms would allow for a prompt isolation. Infection prevention and control in the household is one of the elements that the WHO has identified as critical, due to both the high risk of intra-family transmission and the clinical management of the cases. Isolation in structures such as requisitioned hotels should be considered, where not only those infected and their contacts can be isolated but also infected healthcare workers, thus allowing for care provision.

A task force of the Ministry of Innovation and Digitalisation (MID) is working on solutions for the active tracking of movements through the use of mobile phones and possibly an app that would bring together all three functions: the CFCT, telemedicine and data gathering. Restrictions due the law on confidentiality and individual data protection are a limiting factor in EU, compared to other countries and we will need to work keeping in mind our settings

Research on immunity response to the infection

Various groups are working on developing assays for the assessment of the antibody response. Antibodies, unlike viral RNA detection tests, are indicators of an immune response to infection. They do not indicate whether someone is still contagious and they cannot be used for CFCT, except to trace people who have already had the infection and therefore may be protected. However, there are many aspects which need clarification, such as the sensitivity and specificity of the antibody methods and the duration and extent of the protection conferred by the antibodies measured.

Coordination of the informative systems

In addition to the number of positive results and the number of deaths attributed to COVID-19 that health administrative units gather on a daily basis and send to Istituto Superiore di Sanita’ and Protezione Civile, what has also proved useful is the surveillance of the daily death toll commissioned by the Ministry of Health. The first completed report for the 2019-2020 season was published by the journal Epidemiologia e Prevenzione (3). The system was launched for the monitoring of mortality due to heatwaves and concerned 33 municipalities (all the province capitals and municipalities with over 250,000 inhabitants). As of today only 19 provide up-to-date data. An extension of this system to 1800 municipalities (approximately 20% of all Italian municipalities) has been provided recently.

The integration of all healthcare information systems with the surveillance systems of cases and contacts, would be essential not only to monitor the efficiency and efficacy of the surveillance itself but also to analyse the risk factors for the incidence and lethality of COVID-19. It is worth remembering that a decree which provides for the integration of all the information systems available at national level has already been issued in 2017, but since then there has been no progress in the interconnection of the systems and their operation.

Population sample and estimate of prevalence

Many, including the WHO, have suggested carrying out a statistical survey of the population with different aims: (a) to monitor and study behaviour during the lock-down; (b) to estimate the prevalence of the infection or of immunity. A study using swab testing would offer an estimate of the number of infected at a given moment and strictly depends on the phase of the epidemic in which it is undertaken. It does not seem to be useful at this moment even though it was very informative in Vo’ Euganeo.

There is an increasing interest in seroprevalence studies aimed to identify “immune” subjects, also to be allowed to go back to working settings and possibly re-start economy. However, even the worst case scenario has estimated the number of infected subjects in Italy to be 10% of the entire population. With such low prevalence any further individual decision will be problematic.

If different initiatives are to be undertaken, the Italian Association of Epidemiology is proposing a common “core” protocol for study through serological (antibody) tests, according to the WHO template in order to allow collation of obtained results

We hope that we can reach rapid consensus on these points and that we may pass to the implementation phase.

I thank Francesca Burns, Marina Davoli, Ruggero De Maria and Paola DiGiulio for the exchange of ideas and experiences.