Author: Melis Mevsimler – Ada Lovelace Institute

As the first pandemic of the algorithmic age, the COVID-19 pandemic has brought significant changes to the use of data and artificial intelligence for public health. Governments across the globe saw potential opportunities and benefits in using digital technologies to control the spread of the virus. The result was the development and deployment of novel pandemic technologies.  

As we enter the third year of the pandemic, it is time to investigate whether these technologies have been useful in limiting the spread of COVID-19. This is important retrospectively, but also – and arguably more importantly – because technologies for future pandemics and public health crises will be built on the technical infrastructure, legislation and regulations developed in response to the COVID-19 pandemic.

Contact tracing apps are one of these novel technologies. Installed on an individual’s smartphone, they allow users to report if they are infectious with COVID-19 and record theirinfection status. Using Bluetooth technology, they notify people who have come into close contact with infected individuals  

Manual contact tracing is a well-established public health measure. Contact tracing teams work with infected individuals to recall the people with whom they have had close contact,during the timeframe they may have been infectious. They notify these close contacts and give advice to them on what to do, e.g. to test and self-isolate. 

Digital contact tracing apps were introduced early in the pandemic to complement human contact tracing efforts. At a time when policymakers were looking for effective means to exit from more economically damaging lockdowns, there were high expectations that apps couldidentify chains of infections faster than human tracing teams. The UK Government in particularrolled out its NHS contact tracing app at pace, despite a lack of clear evidence to demonstrate its effectiveness

As contact tracing apps are novel technologies, there is no established evaluation framework for assessing them. Current scientific debate around contact tracing apps largely examinesefficacy by focusing on their technical infrastructure and/or the number of downloads or users. This gives a limited picture of how effective they have been in practice. 

If we want to fully understand the impact of contact tracing apps on public health, we need to understand how they operate in real-world environments. This will help us identify theirbenefits, shortcomings and unintended consequences; and, therefore, improve their currentand future use. 

Today marks the two-year anniversary of the publication of our rapid evidence review of novel COVID-19 technologies Exit through the App Store. As a responsive review of emerging technical responses to COVID-19, including contact tracing apps, the report demonstrated that the governance structures and social impact of these technologies are as important as their efficacy in identifying potentially infected individuals

In this blog post, we reflect on how contact tracing apps have been implemented in the UK and recommend three actions to improve them: how to define and measure real-world effectiveness with a people-centred approach; how to improve governance and social impact with functional communication strategies and how to design socio-economic interventionsaddressing inequalities. 

Action 1: Establisheffectiveness’ with a people-centred approach 

Assessing the effectiveness of contact tracing apps through the lens of users is a worthwhile approach because apps will not achieve their intended purpose unless people use them in the appropriate way, such as reporting infection status, checking notifications, self-isolating if pinned, etc.  

Currently there are three predominant approaches:

  1. The number of people who have downloaded contact tracing apps. Empirical studies that adopt this approach look at the number of people who have downloaded apps and produce statistical models to estimate the impact. They conclude that contact tracing apps have been effective in the UK – although they do not investigate whether people who have downloaded the apps have actively engaged with them.1
  2. The pace of contact tracing apps. This compares the pace of contact tracing apps to human contact tracing efforts. Some scholars argue that contact tracing apps are effective because they detect potentially infected people faster than human teams.2 However, these studies focus on the technical infrastructure of apps and still fail to explore whether people follow the guidelines after having been alerted by the apps.
  3. The number of infected people who have had been alerted by the apps This is based onthe actual number of infected people who have been pinged by the apps. However contact tracing apps in the UK are built withprivacy-preserving’ infrastructure, which means they do not store  user data. Researchers and public health workers cannot therefore identify the number of infected people pinged by the apps. As a literature review from Algorithm Watch demonstrates, in the absence of adequate data researchers produce complex statistical models based on their own assumptions, which are not replicable across diverse geographies and populations.

In practice, despite the growing body of literature, the evidence on the effectiveness of contact tracing apps is inconclusive. This is largely because there is no agreed definition of what ‘effective’ means and, therefore, what measures are needed to assess effectiveness.  

All three approaches analysed have a common limitation: they do not investigate how contact tracing apps impact human behaviour in real-life situations. It is not possible to measure the effectiveness of contact tracing apps without examining people’s behaviours – e.g. knowing how many people stayed home after having been pinged.

Contact tracing apps only have the potential to be effective if deployed as part of a public-health led trajectory towards positive behavioural change. Having established the importance of behaviour, the question is how can we encourage and support people to use contact tracing apps according to their guidelines?

Evidence shows that people’s behaviours in the context of the pandemic are shaped by  their understanding of, and trust in, scientific and Government advice, as well as social and economic factors (e.g. whether they have the financial security to self-isolate).3

Therefore, we need to consider which interventions could have incentivised the use of contact tracing apps, whether these have been carried out successfully and how they could be improved in the future. 

Action 2: Effectively communicate to gain people’s trust

Improving trust is important for improving the effectiveness of contact tracing apps. For people to actively use these apps in line with guidelines, they should trust in their efficacy, as well as trust the government and institutions who implement them. 

Current evidence reveals the low trust in the technology’s effectiveness. This can be seen in the numbers of people who have never actively used contact tracing apps after initially downloading them In July 2021, BBC News reported that around 50,000 people stopped using Protect Scotland.  Similarly, millions of people may have never technically enabled the NHS COVID-19 app (in England and Wales) despite having downloaded it.4 One reason for people to stop using the apps is losing faith in their effectiveness.5

Trust in those who implement contact tracing apps is as important as the trust in the technology itself. It is not surprising that many people were suspicious about pandemic measures, including contact tracing apps. In the beginning of the pandemic, it was reported that reaching herd immunity was the ultimate objective for public health. However,this was later denied or corrected.. This inefficient government communication hampered efforts to convince the public to use digital tools, not only in the UK but across Europe.6 However, the COVID-19 pandemic has shown that the government should improve its ability to translatescientific evidence into clear public health messages in times of crisis.   

Finally, privacy and data protection concerns are among the most important factors that impact public adoption of contact tracing apps.7 A significant number of people are concerned with the potential risk of surveillance through these digital tools. This raises the question of whether the government has effectively communicated the app’s technical infrastructure – which does not track the users’ location as the UK contact tracing apps use Bluetooth not GPS – and engaged with diverse communities. 

Action 3: Design interventions to mitigate socio-economic and health inequalities

Addressing socioeconomic and health inequalities is important for improving contract tracing apps. Vulnerable and marginalised social groups have been disproportionately impacted by the pandemic. Paradoxically, these groups have been less likely to use contact tracing apps. 

Evidence reveals that ethnic minorities’ adoption of contact tracing apps has been lower than the general population.8 At the same time, we know that ethnic minorities are more likely to be infected and to die from COVID-19.9 Professor Landau, the author of People Count: Contact-Tracing Apps and Public Health, proposes a thought-provoking question: what if we deployed contact tracing apps – and other digital tools that specifically address the needs of marginalised communities?

This means reversing the current approach towards emergency public health technologies. The contact tracing apps were deployed with the idea that existing smartphone technologies–Bluetooth and GPS – could be used to digitise the traditional method of contact tracing.

Building technologies in line with the specific needs of marginalised communities requires a significant perspective shift but is reasonable given that they needed the greatest support throughout the pandemic.  As we have stated elsewhere, the government and developers should have built contact tracing apps with social considerations in mind or produced alternative solutions for minority groups.

Furthermore, the ‘digital divide’ has deepened during the pandemic. 5% of households find it hard to pay their broadband bills and/or smartphone services, and 20% of the population do not have foundational digital skills in the UK.10 11 These people could not use contact tracing apps even if they wanted to. This also applies to socioeconomically disadvantaged people. Many people simply cannot afford to stay home after being alerted by a contact tracing app. 

The ‘pingdemic’ combining ‘ping and pandemic – has become a popular phrase to describe the socioeconomic disruption caused by a contact tracing apps. This led the government to update the NHS COVID-19 app (in England and Wales) to decrease the number of contacts being told to isolate

This again shows that technologies cannot work efficiently in isolation and they must be supported by strong social policies. A report by Nuffield Foundation and Resolution Trust demonstrates that the financial support given by the government covers only a quarter of workers’ earnings. This is why the government should consider addressing the intersections of socioeconomic and health inequalities caused by contact tracing apps and other digital tools in the pandemic.  


We have listed three important actions needed to improve the effectiveness of contact tracing apps in the UK. But we realise this is not an exhaustive list. The COVID-19 pandemic has been a multi-dimensional crisis, and there are many other important issues that should be considered.  

For example, contact tracing apps rely extensively on private technical infrastructure built by companies like Apple and Google, and so there could be risks in future surrounding government reliance on the private sector. 

We need to establish a comprehensive evaluation framework for assessing contact tracing apps and learning lessons for the future. Measuring the effectiveness of these technologies in a way that goes beyond technical infrastructure and understands their role in the broader public health ecosystem would be a good place to start. 


  1. See for example, Wymant, C. (2021). ‘The epidemiological impact of the NHS COVID-19 App’. National Institutes of Health. Available at:
  2. Briers, M., Holmes, C. and Fraser, C. (2021). Demonstrating the impact of the NHS COVID-19 app. Alan Turing Institute. Available at:
  3. Li, T. et al. (2021). ‘What makes people install a COVID-19 contact-tracing app? Understanding the influence of app design and individual difference on contact-tracing app adoption intention.’ Pervasive and Mobile Computing. Available at:
  4. Trendall, S. (2021). ‘Data suggests millions of users have not enabled NHS contact-tracing app’. Public Technology. 30 June 2021. Available at:
  5. Garousi, V. and Cutting, D. (2021). ‘What do users think of the UK’s three COVID-19 contact tracing apps? A comparative analysis.’ National Library of Medicine. Available at:
  6. Dewatripont, M. (2021). ‘Vaccination strategies in the midst of an epidemic’. Centre for Economic Policy Research. Available at:
  7. Ada Lovelace Institute. (2021). Public attitudes to COVID-19, technology and inequality: a tracker. Available at:
  8. Ada Lovelace Institute. (2021). Public attitudes to COVID-19, technology and inequality: a tracker. Available at:
  9. Equality Hub and Race Disparity Unit. (2021). Final report on progress to address COVID-19 health inequalities, Available at:
  10. Ofcom. (2021). Affordability of communication services. Available at:
  11. Lloyds Bank. (2021). Essential Digital Skills. Available at:

This piece has been reposted from the Ada Lovelace Institute blog, with permission and thanks. This blog post was first published on the Institute’s website on 19 April 2022. It demonstrates the Institute’s ongoing work on the health data and Covid-19 tech.