Margherita Russo, Claudia Cardinale Ciccotti, Fabrizio De Alexandris, Antonela Gjinaj, Giovanni Romaniello, Antonio Scatorchia, Giorgio Terranova 02 August 2021
During the COVID-19 pandemic, the initial success of countries in using tracking information to contain the spread of infections1 fuelled the expectation that an effective solution essentially required an appropriate technology, a population willing to be tracked, and a public policy that activated tracing in the contagion containment plan.
That things were more complicated than expected was observed by July/August 2020: Van der Leeuw (2020) addressed the general theme of information to contain the pandemic; Savona (2020) focused on the idea of a technological solution without a social context aligned to it and the need to discuss the main players developing those apps, Apple and Google. The numerous questions raised by citizens about privacy were immediately grasped within the EU, by the European Data Protection Board (European Data Protection Board 2020).2
Building on these premises, we focus on dimensions that enhance the success of Science, technology and innovation (STI) policy that supports the use of individual data and AI technologies, accelerated within a few months towards large-scale adoption.3
Contact tracking apps used in nine countries
We conducted a comparative analysis of France, Germany, Italy, and Spain, which were among the countries in Europe most affected by the virus, and the Republic of Ireland, which created one of the best tracking programmes in Europe.
For non-European countries, Australia, New Zealand, the Republic of Korea, and Russia were identified as countries that used different systems and methods for tracking, with opinions that were far from unanimous. The demographic conditions and socio-economic systems of these countries might have affected the capacity and speed of emergency healthcare, limiting/facilitating the effectiveness and speed of fighting the pandemic.
Table 1 (available to download here) displays a summary of the tracing apps’ technological features and information that we comment on along four main dimensions.4
How contact-tracing apps work
Contact-tracing apps differ primarily in how they manage information and transmit data (Bluetooth, Google/Apple, QR codes). The ‘centralised’ apps consolidate data in a system removed from the peripheral devices; ‘decentralised’ apps adopt the technological infrastructure – developed and made available by Apple and Google – in which repositories and data retention are managed by the smartphones.5
Who are the developers? Identikit of a competence network
Tracking app development has created a network of skills that embraces software developers (both the giants like SAP and small companies such as Webtek), telecommunications companies (such as Orange, in France, and Deutsche Telekom, in Germany), academic researchers in many fields, university spin-offs, and civic hackers.
In the world of software developers, new skills are pooled around young developers with high international mobility, such as Bending Spoons (developer of the app Immuni, in Italy). In academic research, consortia have been created to develop strategic research alliances with private companies. In some cases, the app has been designed for non-profit purposes (Australia and Italy), while in others it is a public contract, even if the terms of the contracts were not found.
Implementation of policies for tracking apps
(a) App selection, resources, management, and regulation of information collected.
The Italian government launched an open call, while France, Germany, and Ireland directly identified the company or coalition of companies in charge of app development. The public resources invested for their use and maintenance are clearly identifiable only for Germany.
Except for the Republic of Korea and Russia, the comparative results highlight similar requirements for the management of personal data.6
(b) Integration of the information collected.
We found that in Australia and Italy, the collected information is integrated with the local health system, while in New Zealand, the Republic of Korea and Russia, the central government manages the information directly, in concert with the health authorities. The levels of integration, where present, are poor.
(c) Communication and information campaign.
We found information only about Italy,7 where the information campaign failed just as infections accelerated, in autumn 2020: the opposition parties declared themselves against the tracking app Immuni, claiming they would not download it.
(d) The citizens’ response.
The percentage of the population who were using the contact tracing apps was about 26.6% in Australia, 26.3% in Ireland, 21.7% in Germany, around 16.2% in Italy, and just 3.3% in France: results that are well below the threshold of 60% of population that is considered effective for their use (Hinch et al. 2020).8
A longitudinal perspective: December 2019–June 2021
The sequence of events unfold along five domains (Figure 1): general information on the pandemic and the first case declared in the nine countries, the lockdowns, the adoption of tracking apps, the launch of the first vaccine trial, and the start of vaccine administrations.
Significant public resources were concentrated in countries that had research capacities and production centres of the pharmaceutical industry, which helped produce results that would have been unthinkable under normal conditions.9
Figure 1 Timeline of events: Activation of contact-tracing apps, lockdowns, launch of vaccine trials, general events
Source: Authors’ elaboration on various sources.10
Lessons learned and further developments of the analysis
Despite many countries supporting the use of tracking apps, there seems to be no evidence that they had an effect on controlling COVID-19. Some conditions may explain the success/failure of those policies.
The paradox of privacy and the social dimension of technology
Distant or culturally unrelated countries have encountered similar difficulties getting their citizens to accept the use of tracking apps (Sussman, 2020). In many cases, political interventions negatively influenced public opinion, often well before the apps became downloadable.
While accepting that their personal data are under the control of internet companies, most citizens seem unamenable to sharing their data for the public interest. We argue that the privacy paradox – addressed in terms of ‘democratic compromises and digital surveillance’ (OECD 2021)11 – deserves special attention with respect to conditions for dialogue, to inform citizens and build a sense of collective interest in objectives that require individual commitment to be fully achieved.
The commitment of citizens must be nurtured in normal times to be effective even in emergencies.12 This requires new narratives (Costa-Font 2021) and new practices that involve the community and civil society as essential components of social transformation, as Bowels and Carlin argued (Bowels and Carlin 2020).
How do policymakers choose a technology of public interest?
Countries adopted different technologies and approaches to select the apps and the network of competence to rely on (Table 1). The technological sovereignty of a country – a central element of public policies, also in Europe (Edler et al. 2020, Darnis 2020, European Commission 2020, VDE 2021) – calls for a reflection on the choices of creating or consolidating internal competencies in countries, an issue that goes beyond this pandemic.
Four systemic dimensions of policy implementation
- The first dimension, which is not technological, concerns compatibility with the mobile devices used for data collection: a technological development can fail to consider the users’ characteristics, such as the population’s income distribution and propensity to consume digital products.13
- Second, missing links between a technology – to tackle a health problem – and the health system hamper effective data transmission and recording. Here, decentralisation of health systems matters, since it might increase differences in absorbing/processing/integrating collected data for health surveillance. This issue deserves special attention if the tracking app is to be used as the country exits the pandemic.
- Third, public communication cannot be separated from the implementation of a policy, if technology adoption is to be successful/effective.
- Last, resources for maintaining and developing digital artefacts are important, while related human labour and material infrastructures are often overlooked by decision-makers.
Multilevel interaction calls for new tools for policy analysis
Not new in the analysis of science, technology and innovation policies,14 a systemic perspective can support more effective policy recommendations, grounded on analyses of multilevel interactions and feedback loops in policy implementation (Geyer and Rihani 2012, Gray 2015, Hjern and Porter 1981, Tenbensel 2015). This analytical focus requires a look from inside: an ethnographic method (Agar 1996, 2006) will help to outline a new, complex analytical framework built on participant observation (Bobbio et al. 2017, Vino 2018). Practising such analysis could enhance learning from the current pandemic to improve future science, technology and innovation policies.
Adnkronos (2020), “App anti-Covid in Europa, quanti le scaricano? I numeri”, Adnkronos, 10 June.
Agar, M (2006), “An ethnography by any other name …”, Forum: Qualitative Social Research 7(4).
Agar, M (1996), The professional stranger: An informal introduction to ethnography, 2nd ed., Academic Press.
Bobbio, L, G Pomatto and S Ravazzi (2017), “Le politiche pubbliche: Problemi, soluzioni, incertezze, conflitti”, Mondadori Università.
Chen, L, Y Huang, S Ouyang and W Xiong (2021), “The data privacy paradox and digital demand”, VoxEU.org, 28 June.
Costa-Font, J (2021), “Social value and incentives for vaccine uptake”, VoxEU.org, 29 June 2021.
Darnis, J P (2020), “A COVID-19 moment for technological sovereignty in Europe?”, IAI Istituto Affari Internazionali, 27 April 2020.
Edler, J, K Blind, R Frietsch, S Kimpeler, H Kroll, C Lerch, T Reiss, et al. (2020), “Technology sovereignty: From demand to concept”, Perspectives-Policy Brief.
European Data Protection Board (2020), “Guidelines 03/2020 on the processing of data concerning health for the purpose of scientific research in the context of the COVID-19 outbreak”, European Data Protection Board, 21 April 2020.
European Commission (2020), “Europe: The keys to sovereignty”, European Commission, 11 September 2020.
Geyer, R, and S Rihani (2012), Complexity and public policy, Routledge.
Gray, B (2015), “A case study of complexity and health policy: Planning for a pandemic”, in R Geyer and P Cairney (eds.), Handbook on Complexity and Public Policy, Edward Elgar Publishing.
Hjern, B, and D O Porter (1981), “Implementation structures: A new unit of administrative analysis”, Organization Studies 2(3): 211–27.
Lu Wang, L, K Lo, Y Chandrasekhar, R Reas, J Yang, D Eide, K Funk, et al. (2020), “CORD-19: The Covid-19 open research dataset”, ArXiv, April.
Observatory of Public Sector Innovation (2020), “Singapore is using ‘contact trackers’ to work out who coronavirus patients have come into contact with”, OECD-EU, Observatory of Public Sector Innovation, 20 March 2020.
O’Neill, P H, T Ryan-Mosley, and B Johnson (2020), “A flood of coronavirus apps are tracking us, now it’s time to keep track of them”, MIT Technology Review, 7 May 2020.
OECD – Organisation for Economic Co-operation and Development (2021), “Enhancing access to research data during crises: Lessons learned from the COVID-19 pandemic”, draft summary note of a GSF-RDA workshop, 23 April.
Pancotto, F, and S Righi (2021), “Reflectivity relates differently to pro sociality in naive and strategic subjects”, Nature Scientific Reports.
Savona, M (2020), “The saga of the Covid-19 contact tracing apps: Lessons for data governance”, SPRU Working Paper Series 2020-10.
Shendruk, A (2020), “South Koreans are using smartphone apps to avoid the novel coronavirus”, Quartz, 29 February 2020.
Sussman, N. (2020) ‘Economic incentives and regulation to increase COVID-19 app effectiveness’, VoxEU.org, 18 December.
Tenbensel, T (2015), “Complexity and health policy”, in R Geyer and P Cairney (eds.), Handbook on Complexity and Public Policy, Edward Elgar Publishing.
van der Leeuw, S (2020), “COVID-19 and the role of information processing”, Global Sustainability 3: e27.
VDE Association for Electrical, Electronic and Information Technologies (2021), “New position paper on technological sovereignty for Germany and Europe”, 12 March 2021.
Vino, A (2018), L’attuazione delle politiche pubbliche: Dalla decisione politica all’efficacia sociale, 1st ed., Carocci editore.
World Health Organization (2020), “Contact tracing in the context of COVID-19”, May 2020.
1 As in the case of Singapore (Observatory of Public Sector Innovation 2020) and Republic of Korea (Shendruk 2020), who followed the World Health Organization recommendation (WHO 2020).
2 With regard to other European disciplines on the protection of personal data, the fundamental principles the Member States must comply with were first identified and then further specified for voluntariness, interoperability, regulatory coverage, clarification of purposes, minimization, transparency, protection, security, temporariness.
3 See details on the research project at https://www.poliinnovazione.unimore.it/1257-2/
4 Details of data sources, by country, are available in Russo et al. (2021), Annex 1, and can be browsed online in the Zotero group “OECD_DEMB-internship on contact tracing apps” (register to login).
5 Refer to Savona (2020) for a summary of these technologies and their privacy implications.
6 All the apps require a user’s explicit permission, and a person can choose not to use it without negative effects; policies are in place to ensure that tracking does not survive the specific use of combatting COVID-19; technology and policies ensure that data is deleted when it is no longer needed for public health purposes; user identification is masked or anonymised; policies exist to ensure that only necessary information is collected; sharing data with external entities is prohibited; government and technology are transparent about what data is acquired, from where, how it is used, and who has access to it.
7 The campaign objectives were to: promote the use of Immuni and contribute to the increase of downloads; inform people about the functions of Immuni, its usefulness, safety, and reliability; promote a sense of personal responsibility and belonging to the national community. The campaign was promoted on TV, press, radio, and social media and lasted four months, divided into three phases: the launch in June, a maintenance phase in July/August and early September 2020, and the third recall at the start of autumn. Coordination of the campaign, both for creativity and planning, was handled by Publicis Groupe (a French multinational also based in Italy), which made teams and resources available completely free of charge and coordinated a real alliance among the media involving Rai, Mediaset, Sky, Apple, Google, Facebook, Mondadori, ItaliaOnline, Il Messaggero, RCS, Gedi Group, public figures, startups, and companies.
8 Data are not available for New Zealand, the Republic of Korea, Russia, and Spain (O’Neill et al. 2020)
9 In the race against the pandemic, an open-access collection of documents relating to COVID-19 and online interrogation tools for digital documents in text format was created from scratch (Lu Wang et al. 2020). The OECD-WPTIP conference “Open data and AI analytics in times of COVID-19: The CORD-19 initiative” (30 November 2020) documented the potential to change mindsets regarding open access, particularly in the contributions by Jerry Sheehan (Deputy Director at the National Library of Medicine – National Institutes of Health), Kathryn Funk (Program Manager for PubMed Central at the US National Library of Medicine), and Sebastian Kohlmeier (Sr. Manager of Program Management and Business Operations at the Allen Institute for AI (AI2)).
10 Can be viewed online at https://www.tiki-toki.com/timeline/entry/1639555/COVID-19/.
11 With respect to the consumption sphere, Chen et al. (2021) refer to the ‘data privacy paradox’ specifically as “a general disconnect between consumers’ self-stated privacy preferences and their actual privacy-seeking behaviour”, a phenomenon that in this paper we consider for the implications beyond the consumption sphere.
12 This is demonstrated by the field experiment by Pancotto and Righi (2021), who explored citizens’ participation in seismic emergency contexts and different social propensities to participate. An alternative way to incentivise individual use of a contact tracing up is to limit users’ freedom of daily life, such as access to shops and malls, and cultural and recreational activities. Such incentives might have a short-term effect but does not address the necessary shift from individual behaviour towards a collective goal: an investment that social institutions (not necessarily the state) should foster to achieve any goal, such as the Sustainable Development Goals.
13 In various countries, potential users could not use the app due to software incompatibilities, such as in Portugal, where almost 10% of the population does not have a compatible device with the reference app (Adnkronos 2020).
14 See for example the OECD–TIP projects on system transformation, knowledge triangle and co-creation: https://www.oecd.org/sti/inno/working-group-on-innovation-and-technology-policy.htm.