Skip to main content

Early Evidence of Effectiveness of Digital Contact Tracing for SARS-CoV-2 in Switzerland

Our take —

This proof-of-concept study, available as a preprint and thus not yet peer reviewed, evaluated uptake and effectiveness of SwissCOVID, a decentralized digital contact tracing application (i.e., app) that maintains individual user privacy, using multiple independent data sources from Switzerland. Results suggest that initial uptake of the app was high (~20% of the total Swiss population) and that individuals sought testing as a result of exposure notifications through the app. However, due to a large amount of missing data and because of the decentralized nature of the app itself, quantifying and ultimately interpreting the app’s effectiveness relative to traditional contact tracing methods was difficult. Overall, the results suggest that decentralized digital contact tracing apps may help supplement routine contact tracing activities, but more data are needed to determine their effectiveness.

Study design

Cross-Sectional, Prospective Cohort, Ecological

Study population and setting

SwissCOVID, a digital contact tracing application (i.e., app) launched by the Switzerland public health authority, was deployed to the general population on June 25, 2020. Authors report early findings on the effectiveness of this Exposure Notification (EN) – based contact tracing system from July 23 to September 10, 2020. This EN framework uses Bluetooth Low Energy (BLE) to determine close proximity of two devices (defined as 1.5 meters) and a single-use validation code (CovidCode) linked to the user’s contagious period provided by the Swiss health authority to notify other users who had spent at least 15 minutes within 1.5 meters of a confirmed COVID-19 case. The app is decentralized, meaning that user data is stored locally on users’ phones and not reported to a central authority, thereby preserving individual privacy.

Given the decentralized nature of the app, multiple sources of data were collected to evaluate its effectiveness. App usage was monitored through the Federal Statistical Office and routine public health surveillance data. The Zurich SARS-COV-2 cohort, consisting of 235 index cases and 185 exposed contacts, recruited from August 7 to September 11, 2020, was also used to assess uptake and effectiveness of the app.

Summary of Main Findings

SwissCOVID was downloaded 2.36 million times by September 10, 2020, with an estimated 1.62 million active apps per day, representing 19% of the Swiss population. Out of the 12,456 reported cases during the study period, 2,447 (20%) were provided CovidCodes for the app, and 1,645 (67%) ultimately entered their diagnosis in the app. 1,695 phone calls were received by the SwissCOVID hotline as a result of an app notification (evidence that exposed contacts reacted to the reception of notification). The authors also used routine public health surveillance data and multiple imputation by chained equations (a statistical approach to account for missing data) to estimate that 65 confirmed COVID-19 cases seeked testing as a result of being notified through the SwissCOVID app. Using these same data, the effectiveness of the app was estimated to be similar to that of traditional contact tracing methods using traditional contact tracing data from South Korea and Taiwan. Lastly, 63% (148/235) of confirmed cases in the Zurich SARS-Cov-2 cohort used the app. Among these cases, 127 (86%) inputted their CovidCode diagnosis into the app. 132 of the cohort’s 185 exposed contacts (71%) used the app, of which only 46 (35%) received notification of exposure through the app.

Study Strengths

In this large national study, the authors use multiple independent data sources to evaluate the effectiveness of a digital contact tracing application.

Limitations

There are several notable limitations to this study. First, the total number of users reached through notifications once a confirmed case logs their CovidCode into the app was unknown, limiting our understanding of the reach of digital contact tracing. Second, calls to the SwissCOVID hotlines might not necessarily reflect the adherence of exposed contacts to quarantine recommendations post-notification. Third, it is unclear why out of the 12,456 confirmed cases, only 2,447 were issued a CovidCode. Fourth, information on the time between testing and receiving and entering a CovidCode was not provided, limiting our understanding of digital contact tracing effectiveness. Fifth, there are limitations to the authors’ inference about app usage using routine testing data: only 63% of confirmed cases were asked why they were getting tested; participants could also have reported the appearance of symptoms as the primary reason for getting tested instead of the app itself; and notified contacts may not necessarily have gotten tested, particularly if they were asymptomatic. Finally, the effectiveness of the app in the study population may not necessarily reflect the larger Swiss population. Information about the types of contacts notified through the app (e.g., household, workplace) was not reported, which would have improved understanding of the app’s reach.

Value added

This is among the first studies to report on the effectiveness of a digital contact tracing application and its potential for mitigating the COVID-19 pandemic.

This review was posted on: 28 October 2020