Safe Blues: A Method for Estimation and Control in the Fight Against COVID-19

Overview

Viral spread is a complicated function of multiple elements including biological properties, preventative measures such as sanitation and masks, the environment, and the level of physical proximity. It is this last element that governments can control through social-distancing directives. However, with a pandemic such as COVID-19, the data is always lagging and biased since the time between a patient being infected with SARS-CoV-2 and being recorded as positive can be a week or more. Consequently there can be a time lag of the order of several weeks between the initiation of a regulatory measure and its observed effect. There is thus a pressing need for real time information on the level of physical proximity while respecting personal privacy.

Safe Blues fills this need, providing real time population-level estimates of the level of physical proximity and near-future projections of the epidemic. Safe Blues strands are safe virtual `virus-like' tokens that are spread using cellular devices and Bluetooth. Real time counts of multiple forms of the tokens are combined with delayed measurements of the actual epidemic. Then using machine-learning techniques the Safe Blues system creates more accurate projections about the current and near-future state of the epidemic.

The Safe Blues protocol and machine learning techniques have been developed together with an experimental minimal viable product, presented as an app on Android devices with a server back-end. We are now preparing for a university-wide experiment of Safe Blues.

A diagram of the operation of Safe Blues.
A diagram of the operation of Safe Blues.
Illustration of Safe Blues on a simulated epidemic. The blue lines represent Safe Blues strands, the red markers are the true numbers of infectives, and the blue line represent a Safe Blues prediction of the state of the epidemic. The simulation assumes that real case numbers are delayed by 15 days, which reflects the reality of COVID-19: a long incubation period and mild symptoms at the start of infection mean that diagnoses are significantly delayed, while Safe Blues data is received in real time. In this way, the Safe Blues framework may provide unique, invaluable visibility into the current state of the epidemic and a powerful tool for early detection of subsequent waves or outbreaks.

Key Links

Emulation of epidemics via Bluetooth-based virtual safe virus spread: experimental setup, software, and data - April 01, 2022 medRxiv

Safe Blues: The case for virtual safe virus spread in the long-term fight against epidemics - March 12, 2021 Cell Press Patterns Paper

The 2021/2022 Campus Experiment at The University of Auckland City Campus

Safe Blues Dashboard

2021 Data

Further Material





If you are in UoA Join the experiment now

Presentations and Podcasts

Selected Media

See also a list of further media.

Code and Data

See the Safe Blues GitHub organization

Current Contributors

Previous Contributors

With thanks to

For inquiries, please contact via e-mail.