The acceptability and uptake of smartphone tracking for COVID-19 in Australia

Abstract

In response to the COVID-19 pandemic, many Governments are instituting mobile tracking technologies to perform rapid contact tracing. However, these technologies are only effective if the public is willing to use them, implying that their perceived public health benefits must outweigh personal concerns over privacy and security. The Australian federal government recently launched the 'COVIDSafe' app, designed to anonymously register nearby contacts. If a contact later identifies as infected with COVID-19, health department officials can rapidly followup with their registered contacts to stop the virus' spread. The current study assessed attitudes towards three tracking technologies (telecommunication network tracking, a government app, and Apple and Google's Bluetooth exposure notification system) in two representative samples of the Australian public prior to the launch of COVIDSafe. We compared these attitudes to usage of the COVIDSafe app after its launch in a further two representative samples of the Australian public. Using Bayesian methods, we find widespread acceptance for all tracking technologies, however, observe a large intention-behaviour gap between people's stated attitudes and actual uptake of the COVIDSafe app. We consider the policy implications of these results for Australia and the world at large.

Fig 1. Australian COVID-19 cases, deaths, and key-dates.

COVID-19 daily cases (blue), deaths (red), and key policy decisions (text) in Australia during the COVID-19 pandemic within the period January 23rd–July 14th, 2020. Collection dates of the current study are highlighted in green and the introduction of key tracking technologies are highlighted in yellow. A record of news sources for this fig are included on our OSF page, osf.io/sw7rq.

Fig 2. Mobile tracking technologies.

Infographic highlighting the differences between telecommunication [9], GPS [10], and Bluetooth tracking [11, 12], and the distinction between centralized and decentralized data storage [8].

Fig 3. Survey design for Samples 1–4.

White boxes depict a block of questions with the number of items displayed on the right. Black boxes display the scenario to which participants were randomly assigned (between-subjects design) and gray boxes illustrate judgments of tracking acceptability. ‘Acceptability with other*’ included a local phone data-storage option for the government app scenario and the ability to opt-out of tracking in the telecommunication scenario. Items not included in the results of this paper are shaded gray.

Fig 4. Ordinal regression mean posterior distributions for items assessing the perceived risk from COVID-19 for each sample.

Black points display mean point estimates and coloured error bars display the 95% highest posterior density interval thereof. Dotted lines indicate the ordinal regression threshold parameters which separate the continuous latent variables into the ordinal response categories made by participants (‘None’ to ‘Extremely’). Non-overlapping intervals (within items) are denoted by horizontal bars above each comparison.

Fig 5. Perceived benefits regression.

Ordinal regression mean posterior distributions for items assessing the perceived benefits from tracking for each scenario. Black points display mean point estimates and coloured error bars display the 95% highest posterior density interval thereof. Dotted lines indicate the ordinal regression threshold parameters which separate the continuous latent variables into the ordinal response categories made by participants (‘None’ to ‘Extremely’). Non-overlapping intervals (within items) are denoted by horizontal bars above each comparison.

Fig 6. Privacy perceptions regression.

Ordinal regression mean posterior distributions for items assessing the privacy perceptions of tracking for each scenario. Black points display mean point estimates and coloured error bars display the 95% highest posterior density interval thereof. Dotted lines indicate the ordinal regression threshold parameters which separate the continuous latent variables into the ordinal response categories made by participants (‘None’ to ‘Extremely’). Non-overlapping intervals (within items) are denoted by horizontal bars above each comparison.

Fig 7. Perceived acceptability and uptake.

Acceptability of each tracking scenario collapsed across samples under various conditions. Error bars are 95% Bayesian credible intervals and non-overlapping intervals (within tracking scenario) are denoted by horizontal bars above each comparison.

Fig 8. Perceived acceptability and uptake by date.

Tracking acceptability and COVIDSafe downloads plotted by the date of data collection. COVIDSafe results are displayed for current app downloads (dashed line) and future app downloads (whole-bar). Error bars are 95% Bayesian credible intervals.

Fig 9. Perceived COVIDSafe technology.

Public perceptions of the tracking technology used by the COVIDSafe app, grouped by whether participants reported having downloaded it. Error bars are 95% Bayesian credible intervals. Participants who had downloaded the app were much more likely to report the correct technology used by the app: Bluetooth tracking.