Tracing the virus: a holistic view
Nuria Oliver, Chief Scientific Adviser at Vodafone Institute, provides a holistic approach to tackle the COVID-19 pandemic beyond the use of mobile apps.
Contact tracing (who has not heard of it at this stage of the pandemic?) is undoubtedly a fundamental pillar in the fight against COVID-19, as it allows us to identify, for each positive case of coronavirus and early on, the people they have been in close contact with and thus they might have infected. However, traditional contact tracing suffers from four major limitations.
First, it depends on people’s memory, who need to remember their close contacts in the last few days.
Second, these contacts must be identifiable, i.e. the person infected by COVID-19 must have their contact information so they can be called by the contact tracers. Therefore, manual contact tracing is not able to identify unknown close contacts, such as those that take place on public transport, in a restaurant, or a nightclub.
Third, it is difficult to do the contact tracing across borders.
And, fourthly, this technique requires a lot of human resources to scale. And as it has been reported, unfortunately, the necessary contact tracing resources have not been hired in many countries worldwide.
Given these limitations, together with the wide adoption of smartphones and the fact that the phones are with us most of the time, it seems logical to think that our beloved mobile devices could be used to assist in contact tracing: if my cellphone could communicate, passively and silently, with the mobile phones of people who are close to me, it could automatically record a list of all the close contacts that I have had, without me having to remember them and also including encounters with unknown people.
Therefore, even though manual contact tracing is the main method for the early detection of potentially infected people, the vast majority of countries in the world have deployed some kind of mobile app to complement this manual process.
Bluetooth-based contact tracing apps
Within the different mobile technological solutions for contact tracing, the use of Bluetooth as a sensor to detect the proximity of other mobile devices is the preferred solution in Europe, mainly because of its privacy advantages when compared to other sensors, such as GPS or cell tower information. Singapore was the first country in the world to use Bluetooth in a contact tracing app during the COVID-19 pandemic with its app called Trace Together, developed by the Government Technology Agency and launched on March 20th, 2020. Months later, we find dozens of examples of apps developed by government agencies in many cases in collaboration with academic institutions.
In addition to the use of Bluetooth, there are different possible architectures for these apps. In Europe, we have extensively discussed the centralized vs decentralized approaches. Finally, most European countries, including Spain, Germany, and Italy, have adopted a decentralized model for their contact tracing app because of offering, in theory, greater privacy guarantees than the centralized approaches. To ensure interoperability between Android devices and iPhones, both Google and Apple have developed the GAEN (Google Apple Exposure Notification) interface, which allows smartphones to exchange information, in principle encrypted and preserving people’s privacy, when a close contact occurs, regardless of the operating system of their mobile phone.
Limitations of today’s apps for contact tracing
The usefulness of mobile phones in the fight against COVID-19 is undeniable. However, existing limitations in contact tracing apps should be addressed before they realize their much-hyped potential. I would highlight four key shortcomings:
- The challenge of adoption, both penetration – how many people use them – and geographic, demographic, and socioeconomic coverage, with the risk that pre-existing demographic and socio-economic gaps will be perpetuated or even magnified. It is estimated that roughly 25% of the smartphones worldwide are not compatible with the low energy Bluetooth protocol that the Bluetooth-based contract-tracing apps need. Moreover, we should ensure that the apps are inclusive and accessible;
- Inherent technical limitations of using Bluetooth for a purpose that it was not designed for (such as interference with other devices, sensitivity to the environment, lack of contextual information, e.g. whether people wear masks or not, whether the encounter takes place indoor vs outside if there is a protective screen, etc.) that can lead to a large number of false positives (i.e. close contacts detected by the app that are not) and false negatives (i.e. close contacts that are not detected by the app, even in people who have the app installed);
- The invasion of privacy (especially on Android phones that need to have Google’s location services enabled and use Google Play Services); and
- The lack of evidence so far from the value of contact tracing apps to help cut the chain of transmission, with the added difficulties to measure such value when using a decentralized model.
We know that to achieve high levels of adoption of any technology, people need to trust it. And trust emerges when three conditions are met: (1) competence, i.e. in this case a demonstration that the apps are indeed detecting the close contacts, and only the close contacts; (2) reliability, i.e. sustained competence over time and (3) honesty and transparency, in terms of e.g. the source code of the apps and the libraries that they use, in terms of the data they capture, in terms of their strengths and limitations, etc. Unfortunately, none of these three conditions are met to build trust.
In addition, for these types of apps to be useful, they need not only to accurately detect the maximum number (ideally all) of close contacts and only the close contacts but also, to be able to communicate to those who have had a close contact – in an empathetic and persuasive way at the same time – the need for them to contact their doctor so (s)he can estimate their risk of contagion, get tested if necessary (assuming there is an availability of tests) and accept the consequences of having been a close contact, that is, be willing or able to self-isolate for at least 10 days.
Because we must not forget that the effectiveness of contact tracing depends on the existence of certain infrastructures and capabilities, beyond the human resources (and apps) necessary to detect and interview each positive case.
Effective contact tracing requires more than apps
Contact tracing teams must have access to modern digital tools so they can quickly and efficiently capture and analyze the graph of all contacts.
The ability to quickly test a potentially large number of suspected cases also needs to be in place.
And, infrastructures so that close contacts of positive cases can self isolate need to be deployed. Unfortunately and according to the covid19impactsurvey.org citizen survey, about 50% of its ~260,000 participants in Spain report not being able to quarantine, if necessary, for various reasons. The care of other people (children, parents) is the most important factor, followed by psychological reasons (psychological inability to confine themselves for two weeks and/or fear of stigmatization, which has doubled since the beginning of the pandemic) and labor-economic factors (financial impossibility and fear of losing their work).
In addition, unlike March, today we have valuable information about the nature of the COVID-19 outbreaks and the chains of transmission. In fact, according to official data in Spain, 80% of outbreaks do not exceed 10 people, the vast majority of whom are known of to each other; 42% of outbreaks are of social origin, linked to family gatherings, private parties, and entertainment venues; 20% are of work-related origin and finally, there are mixed outbreaks (16%) of family origin that then move to other environments. To this, we would need to add outbreaks in geriatrics and nursing homes. This knowledge should be a key element in guiding priorities both contact tracing, app deployment, and COVID-19 pandemic containment measures.
The current COVID-19 pandemic demonstrates something that we already knew: public health problems depend not only on the health system but involve society as a whole. Thus, solutions cannot be simple: they require a holistic approach not just for contact tracing, but in the broader sense, for the management of the pandemic.
Given the magnitude of the problem in terms of its difficulty, complexity, costs, and the need for collaboration of those affected, the only thing that will allow us to implement this necessary holistic approach is by strengthening the governance of the crisis, based on a real and loyal collaboration between the different levels (international, national, regional and local) of public administrations, together with actors from the private sector and civil society.
A holistic approach to fight the virus
Technological, health and social tools and processes must be put in place, intertwined, coordinated, and seeking synergies. The contextualization and complementarity of the various tools – including technology – that we use in the fight against coronavirus are fundamental. We must recognize and accept that we are dealing with a virus that will be present in our society for months and with which we must learn to co-exist in a sustainable way – psychologically, economically, environmentally, and socially.
This holistic approach involves not only the development of a clear vision in collaboration with the different stakeholders but more importantly its urgent execution, including the following three key aspects that create a virtuous circle:
- The availability of specific indicators and high-quality data, captured, updated, and shared systematically and regularly, that allow us to make a diagnosis of where we are, analyze the causes, determine what has worked and what has not worked, and model where we are going, enabling decision-making based on evidence and knowledge, including the decision on whether or not to use an app for contact tracing.
- Investment in the necessary human – health and social personnel, contact tracers, teachers, researchers, … — resources for the scale of the challenge, who also need to have all the data, infrastructure, capabilities, and technologies needed to be able to carry out their work effectively.
- The deployment of public policies and specific processes to address weaknesses in the system, including programs to facilitate quarantine (e.g. immediate fully paid medical leave, food and basic necessities pack, caregivers for children or adults who need them, places where citizens can be isolated if not possible at home, guarantees not to lose their job, psychological support, etc.); communication campaigns to foster a culture in which people do not do social activities (including going to work) if they have the slightest suspicion of being infected with coronavirus; protocols to protect the most vulnerable groups; regulations to minimize the risk of contagion in outbreak-prone places and activities (e.g. food processing plants, nightclubs, family celebrations, …) and a set of specific actions for children, adolescents and young people, who are the demographic suffering most intensely the emotional burden of the pandemic.
Finally, we should never forget the spread of the coronavirus depends on each of us, with our responsible or irresponsible behavior. Let us work together, people and technology, civil society, companies, and public administrations, in the fight against the virus. Our union, without a doubt, is what gives us our strength.