Digital tools against COVID-19: taxonomy, ethical challenges, and navigation aid

The collection and use of data is presented as a key strategic remedy by governments and private actors in response to the COVID-19 pandemic. Across countries and institutions, public health experts and researchers from diverse fields such as epidemiology, virology, evolutionary biology, and social science have pointed out the broad range of insights that can be gained by collecting, analysing, and sharing data from diverse digital sources. These sources include data from telephone towers, mobile phone apps, Bluetooth connections, surveillance video, social media feeds, smart thermometers, credit card records, wearables, and several other devices. In parallel, Apple and Google, two of the world’s largest information technology companies, have unprecedentedly banded together to create application programming interfaces that enable an interoperability between Android and iOS devices using apps from public health authorities, to offer a broader Bluetooth-based exposure notification platform by building this functionality into the underlying platforms.

Although the promise of big data analysis has been widely acknowledged, and governments and researchers around the globe are rushing to unlock its potential, notable technical limitations have also surfaced. These limitations include the accuracy, granularity, and quality of data that vary greatly across the different data sources; the adequacy of computation safeguards; and the interoperability issues and security risks. Simultaneously, notable ethical and legal risks and concerns have been identified that accompany digital disease surveillance and prediction.

Civil rights organisations, data protection authorities, and emerging scholars have highlighted the risk of increased digital surveillance after the pandemic.

These groups have emphasised the need to meet baseline conditions such as lawfulness, necessity, and proportionality in data processing, and the need for social justice and fairness to take precedence despite the urgency of this crisis.

As many public and private sector initiatives aiming to use digital technologies in the fight against COVID-19 emerge, the ensuing debate so far seems to be framed generically in a binary choice between using digital technologies to save lives and respecting individual privacy and civil liberties. However, interdisciplinary research has shown the value of context in managing the societal, legal, and ethical risks of data processing for pandemics that stretch beyond the issue of privacy.

^, 

^, 

^, 

In this Health Policy paper, we seek to contribute to the rapidly evolving debate about the promises and risks of digital public health technologies in response to COVID-19. Rather than a focus on so-called solutionist or instrumentalist approaches (where the focus is on the benefit that the technology itself brings to public health management) to digital public health technologies, we instead focus on public health outcomes, as well as the ethical principles guiding these outcomes.

We offer a typology of the main applications that are in use, and we discuss their respective features, including both application-specific and context-specific risks, cross-sectional issues, and ethical concerns. Finally, we propose a navigation aid for policy makers, recommending steps that should be taken to mitigate risks by engaging in a robust risk–benefit analysis. This aid is derived from the translation of ethical principles from public health and data ethics, and builds upon process-based risk assessment and governance frameworks. Further, this aid can be calibrated to each typological domain to guide different technological platforms and at various phases of the deployment of digital public health technology.

Using an established analytical framework for the creation of categorical variables,

we reviewed the rapidly evolving spectrum of digital public health technologies against COVID-19 and created a multidimensional descriptive typology (figure 1). This typology is based on four main categorical variables: key actors, data types, data source, and model of consent. The concept measured by the typology (called the overarching concept in typological research) is the public health function of a technology; not its physical realisation at the hardware or software level. As a result, this multidimensional model can be put in use to categorise not only tools that have already been deployed but also future and emerging technologies. Our typology identifies four main functional categories of digital public health technologies for pandemic management: proximity and contact tracing, symptom monitoring, quarantine control, and flow modelling.