Since the German Corona-Warn-App launched in June 2020, only limited data was available to evaluate its effectiveness in stopping transmission chains, due both to the decentralized design of the system and its separation from public health authorities. In addition to external surveys, an opt-in data donation for privacy-preserving analytics (PPA) and an event-driven user survey (EDUS) were integrated in 2021.
Since March, app users can opt in to automatic, anonymous, daily submission of status data to a backend server for analysis. This data describes app-calculated risk levels and changes to these. In addition, incidence data sets are submitted (1) for newly identified exposure episodes (exposure windows), (2) when lab test results are obtained in the app, and (3) when users who tested positive submit random identifiers (diagnosis keys) to warn their contacts.
This operational data is complemented by a user survey, triggered when the app notifies of a high-risk exposure. A linked online survey focuses on the user’s perception of this notification, prior behavior, and further intentions. At the end of this survey, respondents can agree to a follow-up survey, which inquires about actual behavior changes and particularly about PCR or rapid antigen tests.
Both instruments, PPA and EDUS, include basic demographic data and have some technical precautions against misuse, e.g. a device check and, in the case of EDUS, a one-time token, which was checked and invalidated when entering the online survey.
PPA-participation has reached 7 million daily submissions, while the initial EDUS has been concluded after two months and 26,000 respondents (14,000 also participating in the follow-up survey). Having proved feasibility and acceptance, EDUS may be used for other evaluations.