Epidemiology: Estimating the impact of the NHS COVID-19 app for England and Wales

An analysis of the NHS COVID-19 contact-tracing app for England and Wales estimates that somewhere between approximately 100,000 and 900,000 SARS-CoV-2 infections, depending on the methodology used, were prevented from October to December 2020. For each app user that tested positive for SARS-CoV-2 and allowed their contacts to be notified, at least one further case of infection is thought to have been averted. The findings, reported in Nature this week, provide evidence to support the continued development and deployment of contact tracing apps, in conjunction with other measures (such as social distancing and face coverings), to control the spread of COVID-19.

The NHS COVID-19 app for England and Wales has been regularly used by at least 16.5 million users (28% of the total population) since it launched in September 2020. App users are notified and instructed to quarantine if they have come into close contact with another app user who has a confirmed COVID-19 diagnosis. To assess the impact of this app on reducing transmission, Christophe Fraser and colleagues perform an analysis of data collected in the first three months after the app was launched.

The app sent approximately 1.7 million exposure notifications between October and December 2020 as a result of 560,000 app users testing positive over the study period. The authors use mathematical modelling to estimate that around 6% of app users that receive notifications of close contact with a positive case will go on to report a positive test themselves; similar rates have been observed in manual tracing. To estimate the epidemiological effects of the app, the authors construct a mathematical model to establish the probability that a notified contact would be infected and quarantine in a timely manner. From this modelling, they predict that around 284,000 cases (range 108,000–450,000) were averted. They also use a statistical analysis to compare areas with similar demographics, geography and baseline infection rates, but differing levels of app uptake, to demonstrate that increased app uptake was associated with fewer cases of COVID-19. The statistical analyses suggest a higher number of averted cases, around 594,000 (range 317,000–914,000) during the study period.

The two estimates differ, but agree within their uncertainty, and both demonstrate the value of the app in reducing transmissions, the authors note. They also find that the average number of contacts traced by the app was more than double that of manual contact tracing, although the two methods could be complementary (as they trace different contacts). They add that for every 1% increase in app users, the number of cases could be reduced by 0.8% (by modelling) or 2.3% (by statistical analysis), further supporting the ongoing use of contact tracing apps as part of a system of non-pharmaceutical interventions.

The full paper is available at: https://www.nature.com/articles/s41586-021-03606-z