Bookmarks
Predicting COVID-19 infection rates from ZOE symptom tracker data
Safe People
Organisation name
University of Oxford
Organisation sector
Academic Institute
Safe Projects
Project ID
1150
Lay summary
How accurately can we predict a positive Covid-19 diagnosis, without a test having be taken? Machine Learning based approaches can -- in theory, and with enough data -- provide diagnoses in scenarios where tests are either unavailable or impractical. Using the Covid-19 Symtom Study (with its millions of anonymised observations) which is hosted on the SAIL platform, we will develop, implement and test a set of algorithms which make this prospect a reality.
Public benefit statement
The anticipated outcome is comprehensive proof of the viability of such approaches, and the outperformance of semi-supervised in comparison to purely supervised methods for this critically important public health problem.
Latest approval date
30/09/2020
Safe Data
Dataset(s) name
Data sensitivity level
Anonymous
Legal basis for provision of data under Article 6
(e) processing is necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller;
Lawful conditions for provision of data under Article 9
(j) processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes in accordance with Article 89(1) based on Union or Member State law which shall be proportionate to the aim pursued, respect the essence of the right to data protection and provide for suitable and specific measures to safeguard the fundamental rights and the interests of the data subject.
Common Law Duty of Confidentiality
Not applicable
National data opt-out applied?
Not applicable
Safe Setting
Access type
TRE